phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-03 09:53:45 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

removed error handling for Chunk and add error handling for maybe_await

Browse source
This commit is contained in:
kimbwook 2025-12-02 17:33:05 +09:00
commit 1de6d49064
No known key found for this signature in database
GPG key ID: 13B032C99CBD373A
3053 changed files with 324713 additions and 738731 deletions
Download patch file Download diff file Expand all files Collapse all files

36
docs/README.md
View file

@ -13,6 +13,42 @@ npm run serve
```
You can open up the docs in your browser at http://localhost:3000
## File Import System
This documentation uses `remark-code-import` to import files directly from the repository, eliminating copy-paste maintenance. Files are automatically embedded during build time.
### Importing Code Files
To import Python code (or any code files) with syntax highlighting, use this syntax in `.mdx` files:
```markdown
```python file=./demo_script.py title="demo_script.py"
```
```
This automatically imports the file content and displays it as a formatted code block with Python syntax highlighting.
**Note:** Paths are relative to the current `.mdx` file location, not the repository root.
### Importing Markdown Files as Content
For importing and rendering markdown files (like CONTRIBUTING.md), use the raw-loader approach:
```jsx
import Contributing from '!!raw-loader!../../../CONTRIBUTING.md';
import ReactMarkdown from 'react-markdown';
<ReactMarkdown>{Contributing}</ReactMarkdown>
```
**Requirements:**
- Install dependencies: `npm install --save-dev raw-loader react-markdown`
**Path Resolution:**
- For `remark-code-import`: Paths are relative to the current `.mdx` file location
- For `raw-loader`: Paths are relative to the current `.mdx` file location
- Use `../` to navigate up directories as needed
## Content
Try out Llama Stack's capabilities through our detailed Jupyter notebooks:

62
docs/docs/api-deprecated/index.mdx Normal file
View file

@ -0,0 +1,62 @@
---
title: Deprecated APIs
description: Legacy APIs that are being phased out
sidebar_label: Deprecated
sidebar_position: 1
---
# Deprecated APIs
This section contains APIs that are being phased out in favor of newer, more standardized implementations. These APIs are maintained for backward compatibility but are not recommended for new projects.
:::warning Deprecation Notice
These APIs are deprecated and will be removed in future versions. Please migrate to the recommended alternatives listed below.
:::
## Migration Guide
When using deprecated APIs, please refer to the migration guides provided for each API to understand how to transition to the supported alternatives.
## Deprecated API List
### Legacy Inference APIs
Some older inference endpoints that have been superseded by the standardized Inference API.
**Migration Path:** Use the [Inference API](../api/) instead.
### Legacy Vector Operations
Older vector database operations that have been replaced by the Vector IO API.
**Migration Path:** Use the [Vector IO API](../api/) instead.
### Legacy File Operations
Older file management endpoints that have been replaced by the Files API.
**Migration Path:** Use the [Files API](../api/) instead.
## Support Timeline
Deprecated APIs will be supported according to the following timeline:
- **Current Version**: Full support with deprecation warnings
- **Next Major Version**: Limited support with migration notices
- **Following Major Version**: Removal of deprecated APIs
## Getting Help
If you need assistance migrating from deprecated APIs:
1. Check the specific migration guides for each API
2. Review the [API Reference](../api/) for current alternatives
3. Consult the [Community Forums](https://github.com/llamastack/llama-stack/discussions) for migration support
4. Open an issue on GitHub for specific migration questions
## Contributing
If you find issues with deprecated APIs or have suggestions for improving the migration process, please contribute by:
1. Opening an issue describing the problem
2. Submitting a pull request with improvements
3. Updating migration documentation
For more information on contributing, see our [Contributing Guide](../contributing/).

128
docs/docs/api-experimental/index.mdx Normal file
View file

@ -0,0 +1,128 @@
---
title: Experimental APIs
description: APIs in development with limited support
sidebar_label: Experimental
sidebar_position: 1
---
# Experimental APIs
This section contains APIs that are currently in development and may have limited support or stability. These APIs are available for testing and feedback but should not be used in production environments.
:::warning Experimental Notice
These APIs are experimental and may change without notice. Use with caution and provide feedback to help improve them.
:::
## Current Experimental APIs
### Batch Inference API
Run inference on a dataset of inputs in batch mode for improved efficiency.
**Status:** In Development
**Provider Support:** Limited
**Use Case:** Large-scale inference operations
**Features:**
- Batch processing of multiple inputs
- Optimized resource utilization
- Progress tracking and monitoring
### Batch Agents API
Run agentic workflows on a dataset of inputs in batch mode.
**Status:** In Development
**Provider Support:** Limited
**Use Case:** Large-scale agent operations
**Features:**
- Batch agent execution
- Parallel processing capabilities
- Result aggregation and analysis
### Synthetic Data Generation API
Generate synthetic data for model development and testing.
**Status:** Early Development
**Provider Support:** Very Limited
**Use Case:** Training data augmentation
**Features:**
- Automated data generation
- Quality control mechanisms
- Customizable generation parameters
### Batches API (OpenAI-compatible)
OpenAI-compatible batch management for inference operations.
**Status:** In Development
**Provider Support:** Limited
**Use Case:** OpenAI batch processing compatibility
**Features:**
- OpenAI batch API compatibility
- Job scheduling and management
- Status tracking and monitoring
## Getting Started with Experimental APIs
### Prerequisites
- Llama Stack server running with experimental features enabled
- Appropriate provider configurations
- Understanding of API limitations
### Configuration
Experimental APIs may require special configuration flags or provider settings. Check the specific API documentation for setup requirements.
### Usage Guidelines
1. **Testing Only**: Use experimental APIs for testing and development only
2. **Monitor Changes**: Watch for updates and breaking changes
3. **Provide Feedback**: Report issues and suggest improvements
4. **Backup Data**: Always backup important data when using experimental features
## Feedback and Contribution
We encourage feedback on experimental APIs to help improve them:
### Reporting Issues
- Use GitHub issues with the "experimental" label
- Include detailed error messages and reproduction steps
- Specify the API version and provider being used
### Feature Requests
- Submit feature requests through GitHub discussions
- Provide use cases and expected behavior
- Consider contributing implementations
### Testing
- Test experimental APIs in your environment
- Report performance issues and optimization opportunities
- Share success stories and use cases
## Migration to Stable APIs
As experimental APIs mature, they will be moved to the stable API section. When this happens:
1. **Announcement**: We'll announce the promotion in release notes
2. **Migration Guide**: Detailed migration instructions will be provided
3. **Deprecation Timeline**: Experimental versions will be deprecated with notice
4. **Support**: Full support will be available for stable versions
## Provider Support
Experimental APIs may have limited provider support. Check the specific API documentation for:
- Supported providers
- Configuration requirements
- Known limitations
- Performance characteristics
## Roadmap
Experimental APIs are part of our ongoing development roadmap:
- **Q1 2024**: Batch Inference API stabilization
- **Q2 2024**: Batch Agents API improvements
- **Q3 2024**: Synthetic Data Generation API expansion
- **Q4 2024**: Batches API full OpenAI compatibility
For the latest updates, follow our [GitHub releases](https://github.com/llamastack/llama-stack/releases) and [roadmap discussions](https://github.com/llamastack/llama-stack/discussions).

287
docs/docs/api-openai/index.mdx Normal file
View file

@ -0,0 +1,287 @@
---
title: OpenAI API Compatibility
description: OpenAI-compatible APIs and features in Llama Stack
sidebar_label: OpenAI Compatibility
sidebar_position: 1
---
# OpenAI API Compatibility
Llama Stack provides comprehensive OpenAI API compatibility, allowing you to use existing OpenAI API clients and tools with Llama Stack providers. This compatibility layer ensures seamless migration and interoperability.
## Overview
OpenAI API compatibility in Llama Stack includes:
- **OpenAI-compatible endpoints** for all major APIs
- **Request/response format compatibility** with OpenAI standards
- **Authentication and authorization** using OpenAI-style API keys
- **Error handling** with OpenAI-compatible error codes and messages
- **Rate limiting** and usage tracking compatible with OpenAI patterns
## Supported OpenAI APIs
### Chat Completions API
OpenAI-compatible chat completions for conversational AI applications.
**Endpoint:** `/v1/chat/completions`
**Compatibility:** Full OpenAI API compatibility
**Providers:** All inference providers
**Features:**
- Message-based conversations
- System prompts and user messages
- Function calling support
- Streaming responses
- Temperature and other parameter controls
### Completions API
OpenAI-compatible text completions for general text generation.
**Endpoint:** `/v1/completions`
**Compatibility:** Full OpenAI API compatibility
**Providers:** All inference providers
**Features:**
- Text completion generation
- Prompt engineering support
- Customizable parameters
- Batch processing capabilities
### Embeddings API
OpenAI-compatible embeddings for vector operations.
**Endpoint:** `/v1/embeddings`
**Compatibility:** Full OpenAI API compatibility
**Providers:** All embedding providers
**Features:**
- Text embedding generation
- Multiple embedding models
- Batch embedding processing
- Vector similarity operations
### Files API
OpenAI-compatible file management for document processing.
**Endpoint:** `/v1/files`
**Compatibility:** Full OpenAI API compatibility
**Providers:** Local Filesystem, S3
**Features:**
- File upload and management
- Document processing
- File metadata tracking
- Secure file access
### Vector Store Files API
OpenAI-compatible vector store file operations for RAG applications.
**Endpoint:** `/v1/vector_stores/{vector_store_id}/files`
**Compatibility:** Full OpenAI API compatibility
**Providers:** FAISS, SQLite-vec, Milvus, ChromaDB, Qdrant, Weaviate, Postgres (PGVector)
**Features:**
- Automatic document processing
- Vector store integration
- File chunking and indexing
- Search and retrieval operations
### Batches API
OpenAI-compatible batch processing for large-scale operations.
**Endpoint:** `/v1/batches`
**Compatibility:** OpenAI API compatibility (experimental)
**Providers:** Limited support
**Features:**
- Batch job creation and management
- Progress tracking
- Result retrieval
- Error handling
## Migration from OpenAI
### Step 1: Update API Endpoint
Change your API endpoint from OpenAI to your Llama Stack server:
```python
# Before (OpenAI)
import openai
client = openai.OpenAI(api_key="your-openai-key")
# After (Llama Stack)
import openai
client = openai.OpenAI(
api_key="your-llama-stack-key",
base_url="http://localhost:8000/v1" # Your Llama Stack server
)
```
### Step 2: Configure Providers
Set up your preferred providers in the Llama Stack configuration:
```yaml
# stack-config.yaml
inference:
providers:
- name: "meta-reference"
type: "inline"
model: "llama-3.1-8b"
```
### Step 3: Test Compatibility
Verify that your existing code works with Llama Stack:
```python
# Test chat completions
response = client.chat.completions.create(
model="llama-3.1-8b",
messages=[
{"role": "user", "content": "Hello, world!"}
]
)
print(response.choices[0].message.content)
```
## Provider-Specific Features
### Meta Reference Provider
- Full OpenAI API compatibility
- Local model execution
- Custom model support
### Remote Providers
- OpenAI API compatibility
- Cloud-based execution
- Scalable infrastructure
### Vector Store Providers
- OpenAI vector store API compatibility
- Automatic document processing
- Advanced search capabilities
## Authentication
Llama Stack supports OpenAI-style authentication:
### API Key Authentication
```python
client = openai.OpenAI(
api_key="your-api-key",
base_url="http://localhost:8000/v1"
)
```
### Environment Variables
```bash
export OPENAI_API_KEY="your-api-key"
export OPENAI_BASE_URL="http://localhost:8000/v1"
```
## Error Handling
Llama Stack provides OpenAI-compatible error responses:
```python
try:
response = client.chat.completions.create(...)
except openai.APIError as e:
print(f"API Error: {e}")
except openai.RateLimitError as e:
print(f"Rate Limit Error: {e}")
except openai.APIConnectionError as e:
print(f"Connection Error: {e}")
```
## Rate Limiting
OpenAI-compatible rate limiting is supported:
- **Requests per minute** limits
- **Tokens per minute** limits
- **Concurrent request** limits
- **Usage tracking** and monitoring
## Monitoring and Observability
Track your API usage with OpenAI-compatible monitoring:
- **Request/response logging**
- **Usage metrics** and analytics
- **Performance monitoring**
- **Error tracking** and alerting
## Best Practices
### 1. Provider Selection
Choose providers based on your requirements:
- **Local development**: Meta Reference, Ollama
- **Production**: Cloud providers (Fireworks, Together, NVIDIA)
- **Specialized use cases**: Custom providers
### 2. Model Configuration
Configure models for optimal performance:
- **Model selection** based on task requirements
- **Parameter tuning** for specific use cases
- **Resource allocation** for performance
### 3. Error Handling
Implement robust error handling:
- **Retry logic** for transient failures
- **Fallback providers** for high availability
- **Monitoring** and alerting for issues
### 4. Security
Follow security best practices:
- **API key management** and rotation
- **Access control** and authorization
- **Data privacy** and compliance
## Implementation Examples
For detailed code examples and implementation guides, see our [OpenAI Implementation Guide](../providers/openai.mdx).
## Known Limitations
### Responses API Limitations
The Responses API is still in active development. For detailed information about current limitations and implementation status, see our [OpenAI Responses API Limitations](../providers/openai_responses_limitations.mdx).
## Troubleshooting
### Common Issues
**Connection Errors**
- Verify server is running
- Check network connectivity
- Validate API endpoint URL
**Authentication Errors**
- Verify API key is correct
- Check key permissions
- Ensure proper authentication headers
**Model Errors**
- Verify model is available
- Check provider configuration
- Validate model parameters
### Getting Help
For OpenAI compatibility issues:
1. **Check Documentation**: Review provider-specific documentation
2. **Community Support**: Ask questions in GitHub discussions
3. **Issue Reporting**: Open GitHub issues for bugs
4. **Professional Support**: Contact support for enterprise issues
## Roadmap
Upcoming OpenAI compatibility features:
- **Enhanced batch processing** support
- **Advanced function calling** capabilities
- **Improved error handling** and diagnostics
- **Performance optimizations** for large-scale deployments
For the latest updates, follow our [GitHub releases](https://github.com/llamastack/llama-stack/releases) and [roadmap discussions](https://github.com/llamastack/llama-stack/discussions).

144
docs/docs/api/index.mdx Normal file
View file

@ -0,0 +1,144 @@
---
title: API Reference
description: Complete reference for Llama Stack APIs
sidebar_label: Overview
sidebar_position: 1
---
# API Reference
Llama Stack provides a comprehensive set of APIs for building generative AI applications. All APIs follow OpenAI-compatible standards and can be used interchangeably across different providers.
## Core APIs
### Inference API
Run inference with Large Language Models (LLMs) and embedding models.
**Supported Providers:**
- Meta Reference (Single Node)
- Ollama (Single Node)
- Fireworks (Hosted)
- Together (Hosted)
- NVIDIA NIM (Hosted and Single Node)
- vLLM (Hosted and Single Node)
- TGI (Hosted and Single Node)
- AWS Bedrock (Hosted)
- Cerebras (Hosted)
- Groq (Hosted)
- SambaNova (Hosted)
- PyTorch ExecuTorch (On-device iOS, Android)
- OpenAI (Hosted)
- Anthropic (Hosted)
- Gemini (Hosted)
- WatsonX (Hosted)
### Agents API
Run multi-step agentic workflows with LLMs, including tool usage, memory (RAG), and complex reasoning.
**Supported Providers:**
- Meta Reference (Single Node)
- Fireworks (Hosted)
- Together (Hosted)
- PyTorch ExecuTorch (On-device iOS)
### Vector IO API
Perform operations on vector stores, including adding documents, searching, and deleting documents.
**Supported Providers:**
- FAISS (Single Node)
- SQLite-Vec (Single Node)
- Chroma (Hosted and Single Node)
- Milvus (Hosted and Single Node)
- Postgres (PGVector) (Hosted and Single Node)
- Weaviate (Hosted)
- Qdrant (Hosted and Single Node)
### Files API (OpenAI-compatible)
Manage file uploads, storage, and retrieval with OpenAI-compatible endpoints.
**Supported Providers:**
- Local Filesystem (Single Node)
- S3 (Hosted)
### Vector Store Files API (OpenAI-compatible)
Integrate file operations with vector stores for automatic document processing and search.
**Supported Providers:**
- FAISS (Single Node)
- SQLite-vec (Single Node)
- Milvus (Single Node)
- ChromaDB (Hosted and Single Node)
- Qdrant (Hosted and Single Node)
- Weaviate (Hosted)
- Postgres (PGVector) (Hosted and Single Node)
### Safety API
Apply safety policies to outputs at a systems level, not just model level.
**Supported Providers:**
- Llama Guard (Depends on Inference Provider)
- Prompt Guard (Single Node)
- Code Scanner (Single Node)
- AWS Bedrock (Hosted)
### Post Training API
Fine-tune models for specific use cases and domains.
**Supported Providers:**
- Meta Reference (Single Node)
- HuggingFace (Single Node)
- TorchTune (Single Node)
- NVIDIA NEMO (Hosted)
### Eval API
Generate outputs and perform scoring to evaluate system performance.
**Supported Providers:**
- Meta Reference (Single Node)
- NVIDIA NEMO (Hosted)
### Telemetry API
Collect telemetry data from the system for monitoring and observability.
**Supported Providers:**
- Meta Reference (Single Node)
### Tool Runtime API
Interact with various tools and protocols to extend LLM capabilities.
**Supported Providers:**
- Brave Search (Hosted)
- RAG Runtime (Single Node)
## API Compatibility
All Llama Stack APIs are designed to be OpenAI-compatible, allowing you to:
- Use existing OpenAI API clients and tools
- Migrate from OpenAI to other providers seamlessly
- Maintain consistent API contracts across different environments
## Getting Started
To get started with Llama Stack APIs:
1. **Choose a Distribution**: Select a pre-configured distribution that matches your environment
2. **Configure Providers**: Set up the providers you want to use for each API
3. **Start the Server**: Launch the Llama Stack server with your configuration
4. **Use the APIs**: Make requests to the API endpoints using your preferred client
For detailed setup instructions, see our [Getting Started Guide](../getting_started/quickstart).
## Provider Details
For complete provider compatibility and setup instructions, see our [Providers Documentation](../providers/).
## API Stability
Llama Stack APIs are organized by stability level:
- **[Stable APIs](./index.mdx)** - Production-ready APIs with full support
- **[Experimental APIs](../api-experimental/)** - APIs in development with limited support
- **[Deprecated APIs](../api-deprecated/)** - Legacy APIs being phased out
## OpenAI Integration
For specific OpenAI API compatibility features, see our [OpenAI Compatibility Guide](../api-openai/).

3
docs/docs/building_applications/index.mdx
View file

@ -35,9 +35,6 @@ Here are the key topics that will help you build effective AI applications:
- **[Telemetry](./telemetry.mdx)** - Monitor and analyze your agents' performance and behavior
- **[Safety](./safety.mdx)** - Implement guardrails and safety measures to ensure responsible AI behavior
### 🎮 **Interactive Development**
- **[Playground](./playground.mdx)** - Interactive environment for testing and developing applications
## Application Patterns
### 🤖 **Conversational Agents**

335
docs/docs/building_applications/playground.mdx
View file

@ -1,298 +1,87 @@
---
title: Llama Stack Playground
description: Interactive interface to explore and experiment with Llama Stack capabilities
title: Admin UI & Chat Playground
description: Web-based admin interface and chat playground for Llama Stack
sidebar_label: Playground
sidebar_position: 10
---
import Tabs from '@theme/Tabs';
import TabItem from '@theme/TabItem';
# Admin UI & Chat Playground
# Llama Stack Playground
The Llama Stack UI provides a comprehensive web-based admin interface for managing your Llama Stack server, with an integrated chat playground for interactive testing. This admin interface is the primary way to monitor, manage, and debug your Llama Stack applications.
:::note[Experimental Feature]
The Llama Stack Playground is currently experimental and subject to change. We welcome feedback and contributions to help improve it.
:::
## Quick Start
The Llama Stack Playground is a simple interface that aims to:
- **Showcase capabilities and concepts** of Llama Stack in an interactive environment
- **Demo end-to-end application code** to help users get started building their own applications
- **Provide a UI** to help users inspect and understand Llama Stack API providers and resources
## Key Features
### Interactive Playground Pages
The playground provides interactive pages for users to explore Llama Stack API capabilities:
#### Chatbot Interface
<video
controls
autoPlay
playsInline
muted
loop
style={{width: '100%'}}
>
<source src="https://github.com/user-attachments/assets/8d2ef802-5812-4a28-96e1-316038c84cbf" type="video/mp4" />
Your browser does not support the video tag.
</video>
<Tabs>
<TabItem value="chat" label="Chat">
**Simple Chat Interface**
- Chat directly with Llama models through an intuitive interface
- Uses the `/chat/completions` streaming API under the hood
- Real-time message streaming for responsive interactions
- Perfect for testing model capabilities and prompt engineering
</TabItem>
<TabItem value="rag" label="RAG Chat">
**Document-Aware Conversations**
- Upload documents to create memory banks
- Chat with a RAG-enabled agent that can query your documents
- Uses Llama Stack's `/agents` API to create and manage RAG sessions
- Ideal for exploring knowledge-enhanced AI applications
</TabItem>
</Tabs>
#### Evaluation Interface
<video
controls
autoPlay
playsInline
muted
loop
style={{width: '100%'}}
>
<source src="https://github.com/user-attachments/assets/6cc1659f-eba4-49ca-a0a5-7c243557b4f5" type="video/mp4" />
Your browser does not support the video tag.
</video>
<Tabs>
<TabItem value="scoring" label="Scoring Evaluations">
**Custom Dataset Evaluation**
- Upload your own evaluation datasets
- Run evaluations using available scoring functions
- Uses Llama Stack's `/scoring` API for flexible evaluation workflows
- Great for testing application performance on custom metrics
</TabItem>
<TabItem value="benchmarks" label="Benchmark Evaluations">
<video
controls
autoPlay
playsInline
muted
loop
style={{width: '100%', marginBottom: '1rem'}}
>
<source src="https://github.com/user-attachments/assets/345845c7-2a2b-4095-960a-9ae40f6a93cf" type="video/mp4" />
Your browser does not support the video tag.
</video>
**Pre-registered Evaluation Tasks**
- Evaluate models or agents on pre-defined tasks
- Uses Llama Stack's `/eval` API for comprehensive evaluation
- Combines datasets and scoring functions for standardized testing
**Setup Requirements:**
Register evaluation datasets and benchmarks first:
Launch the admin UI with:
```bash
# Register evaluation dataset
llama-stack-client datasets register \
--dataset-id "mmlu" \
--provider-id "huggingface" \
--url "https://huggingface.co/datasets/llamastack/evals" \
--metadata '{"path": "llamastack/evals", "name": "evals__mmlu__details", "split": "train"}' \
--schema '{"input_query": {"type": "string"}, "expected_answer": {"type": "string"}, "chat_completion_input": {"type": "string"}}'
# Register benchmark task
llama-stack-client benchmarks register \
--eval-task-id meta-reference-mmlu \
--provider-id meta-reference \
--dataset-id mmlu \
--scoring-functions basic::regex_parser_multiple_choice_answer
npx llama-stack-ui
```
</TabItem>
</Tabs>
Then visit `http://localhost:8322` to access the interface.
#### Inspection Interface
## Admin Interface Features
<video
controls
autoPlay
playsInline
muted
loop
style={{width: '100%'}}
>
<source src="https://github.com/user-attachments/assets/01d52b2d-92af-4e3a-b623-a9b8ba22ba99" type="video/mp4" />
Your browser does not support the video tag.
</video>
The Llama Stack UI is organized into three main sections:
<Tabs>
<TabItem value="providers" label="API Providers">
### 🎯 Create
**Chat Playground** - Interactive testing environment
- Real-time chat interface for testing agents and models
- Multi-turn conversations with tool calling support
- Agent SDK integration (will be migrated to Responses API)
- Custom system prompts and model parameter adjustment
**Provider Management**
- Inspect available Llama Stack API providers
- View provider configurations and capabilities
- Uses the `/providers` API for real-time provider information
- Essential for understanding your deployment's capabilities
### 📊 Manage
**Logs & Resource Management** - Monitor and manage your stack
- **Responses Logs**: View and analyze agent responses and interactions
- **Chat Completions Logs**: Monitor chat completion requests and responses
- **Vector Stores**: Create, manage, and monitor vector databases for RAG workflows
- **Prompts**: Full CRUD operations for prompt templates and management
- **Files**: Forthcoming file management capabilities
</TabItem>
<TabItem value="resources" label="API Resources">
## Key Capabilities for Application Development
**Resource Exploration**
- Inspect Llama Stack API resources including:
- **Models**: Available language models
- **Datasets**: Registered evaluation datasets
- **Memory Banks**: Vector databases and knowledge stores
- **Benchmarks**: Evaluation tasks and scoring functions
- **Shields**: Safety and content moderation tools
- Uses `/<resources>/list` APIs for comprehensive resource visibility
- For detailed information about resources, see [Core Concepts](/docs/concepts)
### Real-time Monitoring
- **Response Tracking**: Monitor all agent responses and tool calls
- **Completion Analysis**: View chat completion performance and patterns
- **Vector Store Activity**: Track RAG operations and document processing
- **Prompt Usage**: Analyze prompt template performance
</TabItem>
</Tabs>
### Resource Management
- **Vector Store CRUD**: Create, update, and delete vector databases
- **Prompt Library**: Organize and version control your prompts
- **File Operations**: Manage documents and assets (forthcoming)
### Interactive Testing
- **Chat Playground**: Test conversational flows before production deployment
- **Agent Prototyping**: Validate agent behaviors and tool integrations
## Development Workflow Integration
The admin UI supports your development lifecycle:
1. **Development**: Use chat playground to prototype and test features
2. **Monitoring**: Track system performance through logs and metrics
3. **Management**: Organize prompts, vector stores, and other resources
4. **Debugging**: Analyze logs to identify and resolve issues
## Architecture Notes
- **Current**: Chat playground uses Agents SDK
- **Future**: Migration to Responses API for improved performance and consistency
- **Admin Focus**: Primary emphasis on monitoring, logging, and resource management
## Getting Started
### Quick Start Guide
1. **Launch the UI**: Run `npx llama-stack-ui`
2. **Explore Logs**: Start with Responses and Chat Completions logs to understand your system activity
3. **Test in Playground**: Use the chat interface to validate your agent configurations
4. **Manage Resources**: Create vector stores and organize prompts through the UI
<Tabs>
<TabItem value="setup" label="Setup">
For detailed setup and configuration, see the [Llama Stack UI documentation](/docs/distributions/llama_stack_ui).
**1. Start the Llama Stack API Server**
## Next Steps
```bash
llama stack list-deps together | xargs -L1 uv pip install
llama stack run together
```
**2. Start the Streamlit UI**
```bash
# Launch the playground interface
uv run --with ".[ui]" streamlit run llama_stack.core/ui/app.py
```
</TabItem>
<TabItem value="usage" label="Usage Tips">
**Making the Most of the Playground:**
- **Start with Chat**: Test basic model interactions and prompt engineering
- **Explore RAG**: Upload sample documents to see knowledge-enhanced responses
- **Try Evaluations**: Use the scoring interface to understand evaluation metrics
- **Inspect Resources**: Check what providers and resources are available
- **Experiment with Settings**: Adjust parameters to see how they affect results
</TabItem>
</Tabs>
### Available Distributions
The playground works with any Llama Stack distribution. Popular options include:
<Tabs>
<TabItem value="together" label="Together AI">
```bash
llama stack list-deps together | xargs -L1 uv pip install
llama stack run together
```
**Features:**
- Cloud-hosted models
- Fast inference
- Multiple model options
</TabItem>
<TabItem value="ollama" label="Ollama (Local)">
```bash
llama stack list-deps ollama | xargs -L1 uv pip install
llama stack run ollama
```
**Features:**
- Local model execution
- Privacy-focused
- No internet required
</TabItem>
<TabItem value="meta-reference" label="Meta Reference">
```bash
llama stack list-deps meta-reference | xargs -L1 uv pip install
llama stack run meta-reference
```
**Features:**
- Reference implementation
- All API features available
- Best for development
</TabItem>
</Tabs>
## Use Cases & Examples
### Educational Use Cases
- **Learning Llama Stack**: Hands-on exploration of API capabilities
- **Prompt Engineering**: Interactive testing of different prompting strategies
- **RAG Experimentation**: Understanding how document retrieval affects responses
- **Evaluation Understanding**: See how different metrics evaluate model performance
### Development Use Cases
- **Prototype Testing**: Quick validation of application concepts
- **API Exploration**: Understanding available endpoints and parameters
- **Integration Planning**: Seeing how different components work together
- **Demo Creation**: Showcasing Llama Stack capabilities to stakeholders
### Research Use Cases
- **Model Comparison**: Side-by-side testing of different models
- **Evaluation Design**: Understanding how scoring functions work
- **Safety Testing**: Exploring shield effectiveness with different inputs
- **Performance Analysis**: Measuring model behavior across different scenarios
## Best Practices
### 🚀 **Getting Started**
- Begin with simple chat interactions to understand basic functionality
- Gradually explore more advanced features like RAG and evaluations
- Use the inspection tools to understand your deployment's capabilities
### 🔧 **Development Workflow**
- Use the playground to prototype before writing application code
- Test different parameter settings interactively
- Validate evaluation approaches before implementing them programmatically
### 📊 **Evaluation & Testing**
- Start with simple scoring functions before trying complex evaluations
- Use the playground to understand evaluation results before automation
- Test safety features with various input types
### 🎯 **Production Preparation**
- Use playground insights to inform your production API usage
- Test edge cases and error conditions interactively
- Validate resource configurations before deployment
## Related Resources
- **[Getting Started Guide](../getting_started/quickstart)** - Complete setup and introduction
- **[Core Concepts](/docs/concepts)** - Understanding Llama Stack fundamentals
- **[Agents](./agent)** - Building intelligent agents
- **[RAG (Retrieval Augmented Generation)](./rag)** - Knowledge-enhanced applications
- **[Evaluations](./evals)** - Comprehensive evaluation framework
- **[API Reference](/docs/api/llama-stack-specification)** - Complete API documentation
- Set up your [first agent](/docs/building_applications/agent)
- Implement [RAG functionality](/docs/building_applications/rag)
- Add [evaluation metrics](/docs/building_applications/evals)
- Configure [safety measures](/docs/building_applications/safety)

1
docs/docs/building_applications/safety.mdx
View file

@ -391,5 +391,4 @@ client.shields.register(
- **[Agents](./agent)** - Integrating safety shields with intelligent agents
- **[Agent Execution Loop](./agent_execution_loop)** - Understanding safety in the execution flow
- **[Evaluations](./evals)** - Evaluating safety shield effectiveness
- **[Telemetry](./telemetry)** - Monitoring safety violations and metrics
- **[Llama Guard Documentation](https://github.com/meta-llama/PurpleLlama/tree/main/Llama-Guard3)** - Advanced safety model details

201
docs/docs/building_applications/telemetry.mdx
View file

@ -10,203 +10,34 @@ import TabItem from '@theme/TabItem';
# Telemetry
The Llama Stack uses OpenTelemetry to provide comprehensive tracing, metrics, and logging capabilities.
The preferred way to instrument Llama Stack is with OpenTelemetry. Llama Stack enriches the data
collected by OpenTelemetry to capture helpful information about the performance and behavior of your
application. Here is an example of how to forward your telemetry to an OTLP collector from Llama Stack:
```sh
export OTEL_EXPORTER_OTLP_ENDPOINT="http://127.0.0.1:4318"
export OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf
export OTEL_SERVICE_NAME="llama-stack-server"
## Automatic Metrics Generation
uv pip install opentelemetry-distro opentelemetry-exporter-otlp
uv run opentelemetry-bootstrap -a requirements | uv pip install --requirement -
Llama Stack automatically generates metrics during inference operations. These metrics are aggregated at the **inference request level** and provide insights into token usage and model performance.
### Available Metrics
The following metrics are automatically generated for each inference request:
| Metric Name | Type | Unit | Description | Labels |
|-------------|------|------|-------------|--------|
| `llama_stack_prompt_tokens_total` | Counter | `tokens` | Number of tokens in the input prompt | `model_id`, `provider_id` |
| `llama_stack_completion_tokens_total` | Counter | `tokens` | Number of tokens in the generated response | `model_id`, `provider_id` |
| `llama_stack_tokens_total` | Counter | `tokens` | Total tokens used (prompt + completion) | `model_id`, `provider_id` |
### Metric Generation Flow
1. **Token Counting**: During inference operations (chat completion, completion, etc.), the system counts tokens in both input prompts and generated responses
2. **Metric Construction**: For each request, `MetricEvent` objects are created with the token counts
3. **Telemetry Logging**: Metrics are sent to the configured telemetry sinks
4. **OpenTelemetry Export**: When OpenTelemetry is enabled, metrics are exposed as standard OpenTelemetry counters
### Metric Aggregation Level
All metrics are generated and aggregated at the **inference request level**. This means:
- Each individual inference request generates its own set of metrics
- Metrics are not pre-aggregated across multiple requests
- Aggregation (sums, averages, etc.) can be performed by your observability tools (Prometheus, Grafana, etc.)
- Each metric includes labels for `model_id` and `provider_id` to enable filtering and grouping
### Example Metric Event
```python
MetricEvent(
trace_id="1234567890abcdef",
span_id="abcdef1234567890",
metric="total_tokens",
value=150,
timestamp=1703123456.789,
unit="tokens",
attributes={
"model_id": "meta-llama/Llama-3.2-3B-Instruct",
"provider_id": "tgi"
},
)
uv run opentelemetry-instrument llama stack run run.yaml
```
## Telemetry Sinks
Choose from multiple sink types based on your observability needs:
### Known issues
<Tabs>
<TabItem value="opentelemetry" label="OpenTelemetry">
Some database instrumentation libraries have a known bug where spans get wrapped twice, or do not get connected to a trace.
To prevent this, you can disable database specific tracing, and rely just on the SQLAlchemy tracing. If you are using
`sqlite3` as your database, for example, you can disable the additional tracing like this:
Send events to an OpenTelemetry Collector for integration with observability platforms:
**Use Cases:**
- Visualizing traces in tools like Jaeger
- Collecting metrics for Prometheus
- Integration with enterprise observability stacks
**Features:**
- Standard OpenTelemetry format
- Compatible with all OpenTelemetry collectors
- Supports both traces and metrics
</TabItem>
<TabItem value="console" label="Console">
Print events to the console for immediate debugging:
**Use Cases:**
- Development and testing
- Quick debugging sessions
- Simple logging without external tools
**Features:**
- Immediate output visibility
- No setup required
- Human-readable format
</TabItem>
</Tabs>
## Configuration
### Meta-Reference Provider
Currently, only the meta-reference provider is implemented. It can be configured to send events to multiple sink types:
```yaml
telemetry:
- provider_id: meta-reference
provider_type: inline::meta-reference
config:
service_name: "llama-stack-service"
sinks: ['console', 'otel_trace', 'otel_metric']
otel_exporter_otlp_endpoint: "http://localhost:4318"
```sh
export OTEL_PYTHON_DISABLED_INSTRUMENTATIONS="sqlite3"
```
### Environment Variables
Configure telemetry behavior using environment variables:
- **`OTEL_EXPORTER_OTLP_ENDPOINT`**: OpenTelemetry Collector endpoint (default: `http://localhost:4318`)
- **`OTEL_SERVICE_NAME`**: Service name for telemetry (default: empty string)
- **`TELEMETRY_SINKS`**: Comma-separated list of sinks (default: `[]`)
### Quick Setup: Complete Telemetry Stack
Use the automated setup script to launch the complete telemetry stack (Jaeger, OpenTelemetry Collector, Prometheus, and Grafana):
```bash
./scripts/telemetry/setup_telemetry.sh
```
This sets up:
- **Jaeger UI**: http://localhost:16686 (traces visualization)
- **Prometheus**: http://localhost:9090 (metrics)
- **Grafana**: http://localhost:3000 (dashboards with auto-configured data sources)
- **OTEL Collector**: http://localhost:4318 (OTLP endpoint)
Once running, you can visualize traces by navigating to [Grafana](http://localhost:3000/) and login with login `admin` and password `admin`.
## Querying Metrics
When using the OpenTelemetry sink, metrics are exposed in standard format and can be queried through various tools:
<Tabs>
<TabItem value="prometheus" label="Prometheus Queries">
Example Prometheus queries for analyzing token usage:
```promql
# Total tokens used across all models
sum(llama_stack_tokens_total)
# Tokens per model
sum by (model_id) (llama_stack_tokens_total)
# Average tokens per request over 5 minutes
rate(llama_stack_tokens_total[5m])
# Token usage by provider
sum by (provider_id) (llama_stack_tokens_total)
```
</TabItem>
<TabItem value="grafana" label="Grafana Dashboards">
Create dashboards using Prometheus as a data source:
- **Token Usage Over Time**: Line charts showing token consumption trends
- **Model Performance**: Comparison of different models by token efficiency
- **Provider Analysis**: Breakdown of usage across different providers
- **Request Patterns**: Understanding peak usage times and patterns
</TabItem>
<TabItem value="otlp" label="OpenTelemetry Collector">
Forward metrics to other observability systems:
- Export to multiple backends simultaneously
- Apply transformations and filtering
- Integrate with existing monitoring infrastructure
</TabItem>
</Tabs>
## Best Practices
### 🔍 **Monitoring Strategy**
- Use OpenTelemetry for production environments
- Set up alerts on key metrics like token usage and error rates
### 📊 **Metrics Analysis**
- Track token usage trends to optimize costs
- Monitor response times across different models
- Analyze usage patterns to improve resource allocation
### 🚨 **Alerting & Debugging**
- Set up alerts for unusual token consumption spikes
- Use trace data to debug performance issues
- Monitor error rates and failure patterns
### 🔧 **Configuration Management**
- Use environment variables for flexible deployment
- Ensure proper network access to OpenTelemetry collectors
## Related Resources
- **[Agents](./agent)** - Monitoring agent execution with telemetry
- **[Evaluations](./evals)** - Using telemetry data for performance evaluation
- **[Getting Started Notebook](https://github.com/meta-llama/llama-stack/blob/main/docs/getting_started.ipynb)** - Telemetry examples and queries
- **[OpenTelemetry Documentation](https://opentelemetry.io/)** - Comprehensive observability framework
- **[Jaeger Documentation](https://www.jaegertracing.io/)** - Distributed tracing visualization

22
docs/docs/building_applications/tools.mdx
View file

@ -104,23 +104,19 @@ client.toolgroups.register(
)
```
Note that most of the more useful MCP servers need you to authenticate with them. Many of them use OAuth2.0 for authentication. You can provide authorization headers to send to the MCP server using the "Provider Data" abstraction provided by Llama Stack. When making an agent call,
Note that most of the more useful MCP servers need you to authenticate with them. Many of them use OAuth2.0 for authentication. You can provide the authorization token when creating the Agent:
```python
agent = Agent(
...,
tools=["mcp::deepwiki"],
extra_headers={
"X-LlamaStack-Provider-Data": json.dumps(
{
"mcp_headers": {
"http://mcp.deepwiki.com/sse": {
"Authorization": "Bearer <your_access_token>",
},
},
}
),
},
tools=[
{
"type": "mcp",
"server_url": "https://mcp.deepwiki.com/sse",
"server_label": "mcp::deepwiki",
"authorization": "<your_access_token>", # OAuth token (without "Bearer " prefix)
}
],
)
agent.create_turn(...)
```

9
docs/docs/concepts/apis/external.mdx
View file

@ -58,7 +58,7 @@ External APIs must expose a `available_providers()` function in their module tha
```python
# llama_stack_api_weather/api.py
from llama_stack.providers.datatypes import Api, InlineProviderSpec, ProviderSpec
from llama_stack_api import Api, InlineProviderSpec, ProviderSpec
def available_providers() -> list[ProviderSpec]:
@ -79,7 +79,7 @@ A Protocol class like so:
# llama_stack_api_weather/api.py
from typing import Protocol
from llama_stack.schema_utils import webmethod
from llama_stack_api import webmethod
class WeatherAPI(Protocol):
@ -151,13 +151,12 @@ __all__ = ["WeatherAPI", "available_providers"]
# llama-stack-api-weather/src/llama_stack_api_weather/weather.py
from typing import Protocol
from llama_stack.providers.datatypes import (
from llama_stack_api import (
Api,
ProviderSpec,
RemoteProviderSpec,
webmethod,
)
from llama_stack.schema_utils import webmethod
def available_providers() -> list[ProviderSpec]:
return [

20
docs/docs/concepts/apis/index.mdx
View file

@ -7,7 +7,7 @@ sidebar_position: 1
# APIs
A Llama Stack API is described as a collection of REST endpoints. We currently support the following APIs:
A Llama Stack API is described as a collection of REST endpoints following OpenAI API standards. We currently support the following APIs:
- **Inference**: run inference with a LLM
- **Safety**: apply safety policies to the output at a Systems (not only model) level
@ -16,13 +16,25 @@ A Llama Stack API is described as a collection of REST endpoints. We currently s
- **Scoring**: evaluate outputs of the system
- **Eval**: generate outputs (via Inference or Agents) and perform scoring
- **VectorIO**: perform operations on vector stores, such as adding documents, searching, and deleting documents
- **Telemetry**: collect telemetry data from the system
- **Files**: manage file uploads, storage, and retrieval
- **Post Training**: fine-tune a model
- **Tool Runtime**: interact with various tools and protocols
- **Responses**: generate responses from an LLM using this OpenAI compatible API.
- **Responses**: generate responses from an LLM
We are working on adding a few more APIs to complete the application lifecycle. These will include:
- **Batch Inference**: run inference on a dataset of inputs
- **Batch Agents**: run agents on a dataset of inputs
- **Synthetic Data Generation**: generate synthetic data for model development
- **Batches**: OpenAI-compatible batch management for inference
## OpenAI API Compatibility
We are working on adding OpenAI API compatibility to Llama Stack. This will allow you to use Llama Stack with OpenAI API clients and tools.
### File Operations and Vector Store Integration
The Files API and Vector Store APIs work together through file operations, enabling automatic document processing and search. This integration implements the [OpenAI Vector Store Files API specification](https://platform.openai.com/docs/api-reference/vector-stores-files) and allows you to:
- Upload documents through the Files API
- Automatically process and chunk documents into searchable vectors
- Store processed content in vector databases based on the availability of [our providers](../../providers/index.mdx)
- Search through documents using natural language queries
For detailed information about this integration, see [File Operations and Vector Store Integration](../file_operations_vector_stores.md).

420
docs/docs/concepts/file_operations_vector_stores.mdx Normal file
View file

@ -0,0 +1,420 @@
# File Operations and Vector Store Integration
## Overview
Llama Stack provides seamless integration between the Files API and Vector Store APIs, enabling you to upload documents and automatically process them into searchable vector embeddings. This integration implements file operations following the [OpenAI Vector Store Files API specification](https://platform.openai.com/docs/api-reference/vector-stores-files).
## Enhanced Capabilities Beyond OpenAI
While Llama Stack maintains full compatibility with OpenAI's Vector Store API, it provides several additional capabilities that enhance functionality and flexibility:
### **Embedding Model Specification**
Unlike OpenAI's vector stores which use a fixed embedding model, Llama Stack allows you to specify which embedding model to use when creating a vector store:
```python
# Create vector store with specific embedding model
vector_store = client.vector_stores.create(
name="my_documents",
embedding_model="all-MiniLM-L6-v2", # Specify your preferred model
embedding_dimension=384,
)
```
### **Advanced Search Modes**
Llama Stack supports multiple search modes beyond basic vector similarity:
- **Vector Search**: Pure semantic similarity search using embeddings
- **Keyword Search**: Traditional keyword-based search for exact matches
- **Hybrid Search**: Combines both vector and keyword search for optimal results
```python
# Different search modes
results = await client.vector_stores.search(
vector_store_id=vector_store.id,
query="machine learning algorithms",
search_mode="hybrid", # or "vector", "keyword"
max_num_results=5,
)
```
### **Flexible Ranking Options**
For hybrid search, Llama Stack offers configurable ranking strategies:
- **RRF (Reciprocal Rank Fusion)**: Combines rankings with configurable impact factor
- **Weighted Ranker**: Linear combination of vector and keyword scores with adjustable weights
```python
# Custom ranking configuration
results = await client.vector_stores.search(
vector_store_id=vector_store.id,
query="neural networks",
search_mode="hybrid",
ranking_options={
"ranker": {"type": "weighted", "alpha": 0.7} # 70% vector, 30% keyword
},
)
```
### **Provider Selection**
Choose from multiple vector store providers based on your specific needs:
- **Inline Providers**: FAISS (fast in-memory), SQLite-vec (disk-based), Milvus (high-performance)
- **Remote Providers**: ChromaDB, Qdrant, Weaviate, Postgres (PGVector), Milvus
```python
# Specify provider when creating vector store
vector_store = client.vector_stores.create(
name="my_documents", provider_id="sqlite-vec" # Choose your preferred provider
)
```
## How It Works
The file operations work through several key components:
1. **File Upload**: Documents are uploaded through the Files API
2. **Automatic Processing**: Files are automatically chunked and converted to embeddings
3. **Vector Storage**: Chunks are stored in vector databases with metadata
4. **Search & Retrieval**: Users can search through processed documents using natural language
## Supported Vector Store Providers
The following vector store providers support file operations:
### Inline Providers (Single Node)
- **FAISS**: Fast in-memory vector similarity search
- **SQLite-vec**: Disk-based storage with hybrid search capabilities
### Remote Providers (Hosted)
- **ChromaDB**: Vector database with metadata filtering
- **Weaviate**: Vector database with GraphQL interface
- **Postgres (PGVector)**: Vector extensions for PostgreSQL
### Both Inline & Remote Providers
- **Milvus**: High-performance vector database with advanced indexing
- **Qdrant**: Vector similarity search with payload filtering
## File Processing Pipeline
### 1. File Upload
```python
from llama_stack import LlamaStackClient
client = LlamaStackClient("http://localhost:8000")
# Upload a document
with open("document.pdf", "rb") as f:
file_info = await client.files.upload(file=f, purpose="assistants")
```
### 2. Attach to Vector Store
```python
# Create a vector store
vector_store = client.vector_stores.create(name="my_documents")
# Attach the file to the vector store
file_attach_response = await client.vector_stores.files.create(
vector_store_id=vector_store.id, file_id=file_info.id
)
```
### 3. Automatic Processing
The system automatically:
- Detects the file type and extracts text content
- Splits content into chunks (default: 800 tokens with 400 token overlap)
- Generates embeddings for each chunk
- Stores chunks with metadata in the vector store
- Updates file status to "completed"
### 4. Search and Retrieval
```python
# Search through processed documents
search_results = await client.vector_stores.search(
vector_store_id=vector_store.id,
query="What is the main topic discussed?",
max_num_results=5,
)
# Process results
for result in search_results.data:
print(f"Score: {result.score}")
for content in result.content:
print(f"Content: {content.text}")
```
## Supported File Types
The FileResponse system supports various document formats:
- **Text Files**: `.txt`, `.md`, `.rst`
- **Documents**: `.pdf`, `.docx`, `.doc`
- **Code**: `.py`, `.js`, `.java`, `.cpp`, etc.
- **Data**: `.json`, `.csv`, `.xml`
- **Web Content**: HTML files
## Chunking Strategies
### Default Strategy
The default chunking strategy uses:
- **Max Chunk Size**: 800 tokens
- **Overlap**: 400 tokens
- **Method**: Semantic boundary detection
### Custom Chunking
You can customize chunking when attaching files:
```python
from llama_stack.apis.vector_io import VectorStoreChunkingStrategy
# Attach file with custom chunking
file_attach_response = await client.vector_stores.files.create(
vector_store_id=vector_store.id,
file_id=file_info.id,
chunking_strategy=chunking_strategy,
)
```
**Note**: While Llama Stack is OpenAI-compatible, it also supports additional options beyond the standard OpenAI API. When creating vector stores, you can specify custom embedding models and embedding dimensions that will be used when processing chunks from attached files.
## File Management
### List Files in Vector Store
```python
# List all files in a vector store
files = await client.vector_stores.files.list(vector_store_id=vector_store.id)
for file in files:
print(f"File: {file.filename}, Status: {file.status}")
```
### File Status Tracking
Files go through several statuses:
- **in_progress**: File is being processed
- **completed**: File successfully processed and searchable
- **failed**: Processing failed (check `last_error` for details)
- **cancelled**: Processing was cancelled
### Retrieve File Content
```python
# Get chunked content from vector store
content_response = await client.vector_stores.files.retrieve_content(
vector_store_id=vector_store.id, file_id=file_info.id
)
for chunk in content_response.content:
print(f"Chunk {chunk.metadata.get('chunk_index', 0)}: {chunk.text}")
```
## Vector Store Management
### List Vector Stores
Retrieve a paginated list of all vector stores:
```python
# List all vector stores with default pagination
vector_stores = await client.vector_stores.list()
# Custom pagination and ordering
vector_stores = await client.vector_stores.list(
limit=10,
order="asc", # or "desc"
after="vs_12345678", # cursor-based pagination
)
for store in vector_stores.data:
print(f"Store: {store.name}, Files: {store.file_counts.total}")
print(f"Created: {store.created_at}, Status: {store.status}")
```
### Retrieve Vector Store Details
Get detailed information about a specific vector store:
```python
# Get vector store details
store_details = await client.vector_stores.retrieve(vector_store_id="vs_12345678")
print(f"Name: {store_details.name}")
print(f"Status: {store_details.status}")
print(f"File Counts: {store_details.file_counts}")
print(f"Usage: {store_details.usage_bytes} bytes")
print(f"Created: {store_details.created_at}")
print(f"Metadata: {store_details.metadata}")
```
### Update Vector Store
Modify vector store properties such as name, metadata, or expiration settings:
```python
# Update vector store name and metadata
updated_store = await client.vector_stores.update(
vector_store_id="vs_12345678",
name="Updated Document Collection",
metadata={
"description": "Updated collection for research",
"category": "research",
"version": "2.0",
},
)
# Set expiration policy
expired_store = await client.vector_stores.update(
vector_store_id="vs_12345678",
expires_after={"anchor": "last_active_at", "days": 30},
)
print(f"Updated store: {updated_store.name}")
print(f"Last active: {updated_store.last_active_at}")
```
### Delete Vector Store
Remove a vector store and all its associated data:
```python
# Delete a vector store
delete_response = await client.vector_stores.delete(vector_store_id="vs_12345678")
if delete_response.deleted:
print(f"Vector store {delete_response.id} successfully deleted")
else:
print("Failed to delete vector store")
```
**Important Notes:**
- Deleting a vector store removes all files, chunks, and embeddings
- This operation cannot be undone
- The underlying vector database is also cleaned up
- Consider backing up important data before deletion
## Search Capabilities
### Vector Search
Pure similarity search using embeddings:
```python
results = await client.vector_stores.search(
vector_store_id=vector_store.id,
query="machine learning algorithms",
max_num_results=10,
)
```
### Filtered Search
Combine vector search with metadata filtering:
```python
results = await client.vector_stores.search(
vector_store_id=vector_store.id,
query="machine learning algorithms",
filters={"file_type": "pdf", "upload_date": "2024-01-01"},
max_num_results=10,
)
```
### Hybrid Search
[SQLite-vec](../providers/vector_io/inline_sqlite-vec.mdx), [pgvector](../providers/vector_io/remote_pgvector.mdx), and [Milvus](../providers/vector_io/inline_milvus.mdx) support combining vector and keyword search.
## Performance Considerations
> **Note**: For detailed performance optimization strategies, see [Performance Considerations](../providers/files/openai_file_operations_support.md#performance-considerations) in the provider documentation.
**Key Points:**
- **Chunk Size**: 400-600 tokens for precision, 800-1200 for context
- **Storage**: Choose provider based on your performance needs
- **Search**: Optimize for your specific use case
## Error Handling
> **Note**: For comprehensive troubleshooting and error handling, see [Troubleshooting](../providers/files/openai_file_operations_support.md#troubleshooting) in the provider documentation.
**Common Issues:**
- File processing failures (format, size limits)
- Search performance optimization
- Storage and memory issues
## Best Practices
> **Note**: For detailed best practices and recommendations, see [Best Practices](../providers/files/openai_file_operations_support.md#best-practices) in the provider documentation.
**Key Recommendations:**
- File organization and naming conventions
- Chunking strategy optimization
- Metadata and monitoring practices
- Regular cleanup and maintenance
## Integration Examples
### RAG Application
```python
# Build a RAG system with file uploads
async def build_rag_system():
# Create vector store
vector_store = client.vector_stores.create(name="knowledge_base")
# Upload and process documents
documents = ["doc1.pdf", "doc2.pdf", "doc3.pdf"]
for doc in documents:
with open(doc, "rb") as f:
file_info = await client.files.create(file=f, purpose="assistants")
await client.vector_stores.files.create(
vector_store_id=vector_store.id, file_id=file_info.id
)
return vector_store
# Query the RAG system
async def query_rag(vector_store_id, question):
results = await client.vector_stores.search(
vector_store_id=vector_store_id, query=question, max_num_results=5
)
return results
```
### Document Analysis
```python
# Analyze document content through vector search
async def analyze_document(vector_store_id, file_id):
# Get document content
content = await client.vector_stores.files.retrieve_content(
vector_store_id=vector_store_id, file_id=file_id
)
# Search for specific topics
topics = ["introduction", "methodology", "conclusion"]
analysis = {}
for topic in topics:
results = await client.vector_stores.search(
vector_store_id=vector_store_id, query=topic, max_num_results=3
)
analysis[topic] = results.data
return analysis
```
## Next Steps
- Explore the [Files API documentation](../../providers/files/files.mdx) for detailed API reference
- Check [Vector Store Providers](../providers/vector_io/index.mdx) for specific implementation details
- Review [Getting Started](../getting_started/quickstart.mdx) for quick setup instructions

239
docs/docs/contributing/index.mdx
View file

@ -1,232 +1,13 @@
# Contributing to Llama Stack
We want to make contributing to this project as easy and transparent as
possible.
---
title: Contributing
description: Contributing to Llama Stack
sidebar_label: Contributing to Llama Stack
sidebar_position: 3
hide_title: true
---
## Set up your development environment
import Contributing from '!!raw-loader!../../../CONTRIBUTING.md';
import ReactMarkdown from 'react-markdown';
We use [uv](https://github.com/astral-sh/uv) to manage python dependencies and virtual environments.
You can install `uv` by following this [guide](https://docs.astral.sh/uv/getting-started/installation/).
You can install the dependencies by running:
```bash
cd llama-stack
uv sync --group dev
uv pip install -e .
source .venv/bin/activate
```
```{note}
You can use a specific version of Python with `uv` by adding the `--python <version>` flag (e.g. `--python 3.12`).
Otherwise, `uv` will automatically select a Python version according to the `requires-python` section of the `pyproject.toml`.
For more info, see the [uv docs around Python versions](https://docs.astral.sh/uv/concepts/python-versions/).
```
Note that you can create a dotenv file `.env` that includes necessary environment variables:
```
LLAMA_STACK_BASE_URL=http://localhost:8321
LLAMA_STACK_CLIENT_LOG=debug
LLAMA_STACK_PORT=8321
LLAMA_STACK_CONFIG=<provider-name>
TAVILY_SEARCH_API_KEY=
BRAVE_SEARCH_API_KEY=
```
And then use this dotenv file when running client SDK tests via the following:
```bash
uv run --env-file .env -- pytest -v tests/integration/inference/test_text_inference.py --text-model=meta-llama/Llama-3.1-8B-Instruct
```
### Pre-commit Hooks
We use [pre-commit](https://pre-commit.com/) to run linting and formatting checks on your code. You can install the pre-commit hooks by running:
```bash
uv run pre-commit install
```
After that, pre-commit hooks will run automatically before each commit.
Alternatively, if you don't want to install the pre-commit hooks, you can run the checks manually by running:
```bash
uv run pre-commit run --all-files
```
```{caution}
Before pushing your changes, make sure that the pre-commit hooks have passed successfully.
```
## Discussions -> Issues -> Pull Requests
We actively welcome your pull requests. However, please read the following. This is heavily inspired by [Ghostty](https://github.com/ghostty-org/ghostty/blob/main/CONTRIBUTING.md).
If in doubt, please open a [discussion](https://github.com/meta-llama/llama-stack/discussions); we can always convert that to an issue later.
### Issues
We use GitHub issues to track public bugs. Please ensure your description is
clear and has sufficient instructions to be able to reproduce the issue.
Meta has a [bounty program](http://facebook.com/whitehat/info) for the safe
disclosure of security bugs. In those cases, please go through the process
outlined on that page and do not file a public issue.
### Contributor License Agreement ("CLA")
In order to accept your pull request, we need you to submit a CLA. You only need
to do this once to work on any of Meta's open source projects.
Complete your CLA here: [https://code.facebook.com/cla](https://code.facebook.com/cla)
**I'd like to contribute!**
If you are new to the project, start by looking at the issues tagged with "good first issue". If you're interested
leave a comment on the issue and a triager will assign it to you.
Please avoid picking up too many issues at once. This helps you stay focused and ensures that others in the community also have opportunities to contribute.
- Try to work on only 1–2 issues at a time, especially if you’re still getting familiar with the codebase.
- Before taking an issue, check if it’s already assigned or being actively discussed.
- If you’re blocked or can’t continue with an issue, feel free to unassign yourself or leave a comment so others can step in.
**I have a bug!**
1. Search the issue tracker and discussions for similar issues.
2. If you don't have steps to reproduce, open a discussion.
3. If you have steps to reproduce, open an issue.
**I have an idea for a feature!**
1. Open a discussion.
**I've implemented a feature!**
1. If there is an issue for the feature, open a pull request.
2. If there is no issue, open a discussion and link to your branch.
**I have a question!**
1. Open a discussion or use [Discord](https://discord.gg/llama-stack).
**Opening a Pull Request**
1. Fork the repo and create your branch from `main`.
2. If you've changed APIs, update the documentation.
3. Ensure the test suite passes.
4. Make sure your code lints using `pre-commit`.
5. If you haven't already, complete the Contributor License Agreement ("CLA").
6. Ensure your pull request follows the [conventional commits format](https://www.conventionalcommits.org/en/v1.0.0/).
7. Ensure your pull request follows the [coding style](#coding-style).
Please keep pull requests (PRs) small and focused. If you have a large set of changes, consider splitting them into logically grouped, smaller PRs to facilitate review and testing.
```{tip}
As a general guideline:
- Experienced contributors should try to keep no more than 5 open PRs at a time.
- New contributors are encouraged to have only one open PR at a time until they’re familiar with the codebase and process.
```
## Repository guidelines
### Coding Style
* Comments should provide meaningful insights into the code. Avoid filler comments that simply
describe the next step, as they create unnecessary clutter, same goes for docstrings.
* Prefer comments to clarify surprising behavior and/or relationships between parts of the code
rather than explain what the next line of code does.
* Catching exceptions, prefer using a specific exception type rather than a broad catch-all like
`Exception`.
* Error messages should be prefixed with "Failed to ..."
* 4 spaces for indentation rather than tab
* When using `# noqa` to suppress a style or linter warning, include a comment explaining the
justification for bypassing the check.
* When using `# type: ignore` to suppress a mypy warning, include a comment explaining the
justification for bypassing the check.
* Don't use unicode characters in the codebase. ASCII-only is preferred for compatibility or
readability reasons.
* Providers configuration class should be Pydantic Field class. It should have a `description` field
that describes the configuration. These descriptions will be used to generate the provider
documentation.
* When possible, use keyword arguments only when calling functions.
* Llama Stack utilizes custom Exception classes for certain Resources that should be used where applicable.
### License
By contributing to Llama, you agree that your contributions will be licensed
under the LICENSE file in the root directory of this source tree.
## Common Tasks
Some tips about common tasks you work on while contributing to Llama Stack:
### Setup for development
```bash
git clone https://github.com/meta-llama/llama-stack.git
cd llama-stack
uv run llama stack list-deps <distro-name> | xargs -L1 uv pip install
# (Optional) If you are developing the llama-stack-client-python package, you can add it as an editable package.
git clone https://github.com/meta-llama/llama-stack-client-python.git
uv add --editable ../llama-stack-client-python
```
### Updating distribution configurations
If you have made changes to a provider's configuration in any form (introducing a new config key, or
changing models, etc.), you should run `./scripts/distro_codegen.py` to re-generate various YAML
files as well as the documentation. You should not change `docs/source/.../distributions/` files
manually as they are auto-generated.
### Updating the provider documentation
If you have made changes to a provider's configuration, you should run `./scripts/provider_codegen.py`
to re-generate the documentation. You should not change `docs/source/.../providers/` files manually
as they are auto-generated.
Note that the provider "description" field will be used to generate the provider documentation.
### Building the Documentation
If you are making changes to the documentation at [https://llamastack.github.io/](https://llamastack.github.io/), you can use the following command to build the documentation and preview your changes.
```bash
# This rebuilds the documentation pages and the OpenAPI spec.
npm install
npm run gen-api-docs all
npm run build
# This will start a local server (usually at http://127.0.0.1:3000).
npm run serve
```
### Update API Documentation
If you modify or add new API endpoints, update the API documentation accordingly. You can do this by running the following command:
```bash
uv run ./docs/openapi_generator/run_openapi_generator.sh
```
The generated API schema will be available in `docs/static/`. Make sure to review the changes before committing.
## Adding a New Provider
See:
- [Adding a New API Provider Page](./new_api_provider.mdx) which describes how to add new API providers to the Stack.
- [Vector Database Page](./new_vector_database.mdx) which describes how to add a new vector databases with Llama Stack.
- [External Provider Page](/docs/providers/external/) which describes how to add external providers to the Stack.
## Testing
See the [Testing README](https://github.com/meta-llama/llama-stack/blob/main/tests/README.md) for detailed testing information.
## Advanced Topics
For developers who need deeper understanding of the testing system internals:
- [Record-Replay Testing](./testing/record-replay.mdx)
### Benchmarking
See the [Benchmarking README](https://github.com/meta-llama/llama-stack/blob/main/benchmarking/k8s-benchmark/README.md) for benchmarking information.
<ReactMarkdown>{Contributing}</ReactMarkdown>

217
docs/docs/deploying/kubernetes_deployment.mdx
View file

@ -10,7 +10,7 @@ import TabItem from '@theme/TabItem';
# Kubernetes Deployment Guide
Deploy Llama Stack and vLLM servers in a Kubernetes cluster instead of running them locally. This guide covers both local development with Kind and production deployment on AWS EKS.
Deploy Llama Stack and vLLM servers in a Kubernetes cluster instead of running them locally. This guide covers deployment using the Kubernetes operator to manage the Llama Stack server with Kind. The vLLM inference server is deployed manually.
## Prerequisites
@ -110,115 +110,176 @@ spec:
EOF
```
### Step 3: Configure Llama Stack
### Step 3: Install Kubernetes Operator
Update your run configuration:
```yaml
providers:
inference:
- provider_id: vllm
provider_type: remote::vllm
config:
url: http://vllm-server.default.svc.cluster.local:8000/v1
max_tokens: 4096
api_token: fake
```
Build container image:
Install the Llama Stack Kubernetes operator to manage Llama Stack deployments:
```bash
tmp_dir=$(mktemp -d) && cat >$tmp_dir/Containerfile.llama-stack-run-k8s <<EOF
FROM distribution-myenv:dev
RUN apt-get update && apt-get install -y git
RUN git clone https://github.com/meta-llama/llama-stack.git /app/llama-stack-source
ADD ./vllm-llama-stack-run-k8s.yaml /app/config.yaml
EOF
podman build -f $tmp_dir/Containerfile.llama-stack-run-k8s -t llama-stack-run-k8s $tmp_dir
# Install from the latest main branch
kubectl apply -f https://raw.githubusercontent.com/llamastack/llama-stack-k8s-operator/main/release/operator.yaml
# Or install a specific version (e.g., v0.4.0)
# kubectl apply -f https://raw.githubusercontent.com/llamastack/llama-stack-k8s-operator/v0.4.0/release/operator.yaml
```
### Step 4: Deploy Llama Stack Server
Verify the operator is running:
```bash
kubectl get pods -n llama-stack-operator-system
```
For more information about the operator, see the [llama-stack-k8s-operator repository](https://github.com/llamastack/llama-stack-k8s-operator).
### Step 4: Deploy Llama Stack Server using Operator
Create a `LlamaStackDistribution` custom resource to deploy the Llama Stack server. The operator will automatically create the necessary Deployment, Service, and other resources.
You can optionally override the default `run.yaml` using `spec.server.userConfig` with a ConfigMap (see [userConfig spec](https://github.com/llamastack/llama-stack-k8s-operator/blob/main/docs/api-overview.md#userconfigspec)).
```yaml
cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: PersistentVolumeClaim
apiVersion: llamastack.io/v1alpha1
kind: LlamaStackDistribution
metadata:
name: llama-pvc
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 1Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: llama-stack-server
name: llamastack-vllm
spec:
replicas: 1
selector:
matchLabels:
app.kubernetes.io/name: llama-stack
template:
metadata:
labels:
app.kubernetes.io/name: llama-stack
spec:
containers:
- name: llama-stack
image: localhost/llama-stack-run-k8s:latest
imagePullPolicy: IfNotPresent
command: ["llama", "stack", "run", "/app/config.yaml"]
ports:
- containerPort: 5000
volumeMounts:
- name: llama-storage
mountPath: /root/.llama
volumes:
- name: llama-storage
persistentVolumeClaim:
claimName: llama-pvc
---
apiVersion: v1
kind: Service
metadata:
name: llama-stack-service
spec:
selector:
app.kubernetes.io/name: llama-stack
ports:
- protocol: TCP
port: 5000
targetPort: 5000
type: ClusterIP
server:
distribution:
name: starter
containerSpec:
port: 8321
env:
- name: VLLM_URL
value: "http://vllm-server.default.svc.cluster.local:8000/v1"
- name: VLLM_MAX_TOKENS
value: "4096"
- name: VLLM_API_TOKEN
value: "fake"
# Optional: override run.yaml from a ConfigMap using userConfig
userConfig:
configMap:
name: llama-stack-config
storage:
size: "20Gi"
mountPath: "/home/lls/.lls"
EOF
```
**Configuration Options:**
- `replicas`: Number of Llama Stack server instances to run
- `server.distribution.name`: The distribution to use (e.g., `starter` for the starter distribution). See the [list of supported distributions](https://github.com/llamastack/llama-stack-k8s-operator/blob/main/distributions.json) in the operator repository.
- `server.distribution.image`: (Optional) Custom container image for non-supported distributions. Use this field when deploying a distribution that is not in the supported list. If specified, this takes precedence over `name`.
- `server.containerSpec.port`: Port on which the Llama Stack server listens (default: 8321)
- `server.containerSpec.env`: Environment variables to configure providers:
- `server.userConfig`: (Optional) Override the default `run.yaml` using a ConfigMap. See [userConfig spec](https://github.com/llamastack/llama-stack-k8s-operator/blob/main/docs/api-overview.md#userconfigspec).
- `server.storage.size`: Size of the persistent volume for model and data storage
- `server.storage.mountPath`: Where to mount the storage in the container
**Note:** For a complete list of supported distributions, see [distributions.json](https://github.com/llamastack/llama-stack-k8s-operator/blob/main/distributions.json) in the operator repository. To use a custom or non-supported distribution, set the `server.distribution.image` field with your container image instead of `server.distribution.name`.
The operator automatically creates:
- A Deployment for the Llama Stack server
- A Service to access the server
- A PersistentVolumeClaim for storage
- All necessary RBAC resources
Check the status of your deployment:
```bash
kubectl get llamastackdistribution
kubectl describe llamastackdistribution llamastack-vllm
```
### Step 5: Test Deployment
Wait for the Llama Stack server pod to be ready:
```bash
# Port forward and test
kubectl port-forward service/llama-stack-service 5000:5000
llama-stack-client --endpoint http://localhost:5000 inference chat-completion --message "hello, what model are you?"
# Check the status of the LlamaStackDistribution
kubectl get llamastackdistribution llamastack-vllm
# Check the pods created by the operator
kubectl get pods -l app.kubernetes.io/name=llama-stack
# Wait for the pod to be ready
kubectl wait --for=condition=ready pod -l app.kubernetes.io/name=llama-stack --timeout=300s
```
Get the service name created by the operator (it typically follows the pattern `<llamastackdistribution-name>-service`):
```bash
# List services to find the service name
kubectl get services | grep llamastack
# Port forward and test (replace SERVICE_NAME with the actual service name)
kubectl port-forward service/llamastack-vllm-service 8321:8321
```
In another terminal, test the deployment:
```bash
llama-stack-client --endpoint http://localhost:8321 inference chat-completion --message "hello, what model are you?"
```
## Troubleshooting
**Check pod status:**
### vLLM Server Issues
**Check vLLM pod status:**
```bash
kubectl get pods -l app.kubernetes.io/name=vllm
kubectl logs -l app.kubernetes.io/name=vllm
```
**Test service connectivity:**
**Test vLLM service connectivity:**
```bash
kubectl run -it --rm debug --image=curlimages/curl --restart=Never -- curl http://vllm-server:8000/v1/models
```
### Llama Stack Server Issues
**Check LlamaStackDistribution status:**
```bash
# Get detailed status
kubectl describe llamastackdistribution llamastack-vllm
# Check for events
kubectl get events --sort-by='.lastTimestamp' | grep llamastack-vllm
```
**Check operator-managed pods:**
```bash
# List all pods managed by the operator
kubectl get pods -l app.kubernetes.io/name=llama-stack
# Check pod logs (replace POD_NAME with actual pod name)
kubectl logs -l app.kubernetes.io/name=llama-stack
```
**Check operator status:**
```bash
# Verify the operator is running
kubectl get pods -n llama-stack-operator-system
# Check operator logs if issues persist
kubectl logs -n llama-stack-operator-system -l control-plane=controller-manager
```
**Verify service connectivity:**
```bash
# Get the service endpoint
kubectl get svc llamastack-vllm-service
# Test connectivity from within the cluster
kubectl run -it --rm debug --image=curlimages/curl --restart=Never -- curl http://llamastack-vllm-service:8321/health
```
## Related Resources
- **[Deployment Overview](/docs/deploying/)** - Overview of deployment options
- **[Distributions](/docs/distributions)** - Understanding Llama Stack distributions
- **[Configuration](/docs/distributions/configuration)** - Detailed configuration options
- **[LlamaStack Operator](https://github.com/llamastack/llama-stack-k8s-operator)** - Overview of llama-stack kubernetes operator
- **[LlamaStackDistribution](https://github.com/llamastack/llama-stack-k8s-operator/blob/main/docs/api-overview.md)** - API Spec of the llama-stack operator Custom Resource.

2
docs/docs/distributions/building_distro.mdx
View file

@ -65,7 +65,7 @@ external_providers_dir: /workspace/providers.d
Inside `providers.d/custom_ollama/provider.py`, define `get_provider_spec()` so the CLI can discover dependencies:
```python
from llama_stack.providers.datatypes import ProviderSpec
from llama_stack_api.providers.datatypes import ProviderSpec
def get_provider_spec() -> ProviderSpec:

41
docs/docs/distributions/configuration.mdx
View file

@ -21,7 +21,6 @@ apis:
- inference
- vector_io
- safety
- telemetry
providers:
inference:
- provider_id: ollama
@ -51,10 +50,6 @@ providers:
responses:
backend: sql_default
table_name: responses
telemetry:
- provider_id: meta-reference
provider_type: inline::meta-reference
config: {}
storage:
backends:
kv_default:
@ -63,13 +58,21 @@ storage:
sql_default:
type: sql_sqlite
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/ollama}/sqlstore.db
references:
stores:
metadata:
backend: kv_default
namespace: registry
inference:
backend: sql_default
table_name: inference_store
max_write_queue_size: 10000
num_writers: 4
conversations:
backend: sql_default
table_name: openai_conversations
prompts:
backend: kv_default
namespace: prompts
models:
- metadata: {}
model_id: ${env.INFERENCE_MODEL}
@ -92,7 +95,6 @@ apis:
- inference
- vector_io
- safety
- telemetry
```
## Providers
@ -219,7 +221,15 @@ models:
```
A Model is an instance of a "Resource" (see [Concepts](../concepts/)) and is associated with a specific inference provider (in this case, the provider with identifier `ollama`). This is an instance of a "pre-registered" model. While we always encourage the clients to register models before using them, some Stack servers may come up a list of "already known and available" models.
What's with the `provider_model_id` field? This is an identifier for the model inside the provider's model catalog. Contrast it with `model_id` which is the identifier for the same model for Llama Stack's purposes. For example, you may want to name "llama3.2:vision-11b" as "image_captioning_model" when you use it in your Stack interactions. When omitted, the server will set `provider_model_id` to be the same as `model_id`.
What's with the `provider_model_id` field? This is an identifier for the model inside the provider's model catalog. The `model_id` field is provided for configuration purposes but is not used as part of the model identifier.
**Important:** Models are identified as `provider_id/provider_model_id` in the system and when making API calls. When `provider_model_id` is omitted, the server will set it to be the same as `model_id`.
Examples:
- Config: `model_id: llama3.2`, `provider_id: ollama`, `provider_model_id: null`
→ Access as: `ollama/llama3.2`
- Config: `model_id: my-llama`, `provider_id: vllm-inference`, `provider_model_id: llama-3-2-3b`
→ Access as: `vllm-inference/llama-3-2-3b` (the `model_id` is not used in the identifier)
If you need to conditionally register a model in the configuration, such as only when specific environment variable(s) are set, this can be accomplished by utilizing a special `__disabled__` string as the default value of an environment variable substitution, as shown below:
@ -589,24 +599,13 @@ created by users sharing a team with them:
In addition to resource-based access control, Llama Stack supports endpoint-level authorization using OAuth 2.0 style scopes. When authentication is enabled, specific API endpoints require users to have particular scopes in their authentication token.
**Scope-Gated APIs:**
The following APIs are currently gated by scopes:
- **Telemetry API** (scope: `telemetry.read`):
- `POST /telemetry/traces` - Query traces
- `GET /telemetry/traces/{trace_id}` - Get trace by ID
- `GET /telemetry/traces/{trace_id}/spans/{span_id}` - Get span by ID
- `POST /telemetry/spans/{span_id}/tree` - Get span tree
- `POST /telemetry/spans` - Query spans
- `POST /telemetry/metrics/{metric_name}` - Query metrics
**Authentication Configuration:**
For **JWT/OAuth2 providers**, scopes should be included in the JWT's claims:
```json
{
"sub": "user123",
"scope": "telemetry.read",
"scope": "<scope>",
"aud": "llama-stack"
}
```
@ -616,7 +615,7 @@ For **custom authentication providers**, the endpoint must return user attribute
{
"principal": "user123",
"attributes": {
"scopes": ["telemetry.read"]
"scopes": ["<scope>"]
}
}
```

2
docs/docs/distributions/importing_as_library.mdx
View file

@ -11,7 +11,7 @@ If you are planning to use an external service for Inference (even Ollama or TGI
This avoids the overhead of setting up a server.
```bash
# setup
uv pip install llama-stack
uv pip install llama-stack llama-stack-client
llama stack list-deps starter | xargs -L1 uv pip install
```

1
docs/docs/distributions/index.mdx
View file

@ -19,3 +19,4 @@ This section provides an overview of the distributions available in Llama Stack.
- **[Starting Llama Stack Server](./starting_llama_stack_server.mdx)** - How to run distributions
- **[Importing as Library](./importing_as_library.mdx)** - Use distributions in your code
- **[Configuration Reference](./configuration.mdx)** - Configuration file format details
- **[Llama Stack UI](./llama_stack_ui.mdx)** - Web-based user interface for interacting with Llama Stack servers

17
docs/docs/distributions/k8s/stack-configmap.yaml
View file

@ -8,7 +8,6 @@ data:
- inference
- files
- safety
- telemetry
- tool_runtime
- vector_io
providers:
@ -73,12 +72,6 @@ data:
db: ${env.POSTGRES_DB:=llamastack}
user: ${env.POSTGRES_USER:=llamastack}
password: ${env.POSTGRES_PASSWORD:=llamastack}
telemetry:
- provider_id: meta-reference
provider_type: inline::meta-reference
config:
service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
sinks: ${env.TELEMETRY_SINKS:=console}
tool_runtime:
- provider_id: brave-search
provider_type: remote::brave-search
@ -113,13 +106,21 @@ data:
db: ${env.POSTGRES_DB:=llamastack}
user: ${env.POSTGRES_USER:=llamastack}
password: ${env.POSTGRES_PASSWORD:=llamastack}
references:
stores:
metadata:
backend: kv_default
namespace: registry
inference:
backend: sql_default
table_name: inference_store
max_write_queue_size: 10000
num_writers: 4
conversations:
backend: sql_default
table_name: openai_conversations
prompts:
backend: kv_default
namespace: prompts
models:
- metadata:
embedding_dimension: 768

114
docs/docs/distributions/k8s/stack_run_config.yaml
View file

@ -5,7 +5,6 @@ apis:
- inference
- files
- safety
- telemetry
- tool_runtime
- vector_io
providers:
@ -32,21 +31,17 @@ providers:
provider_type: remote::chromadb
config:
url: ${env.CHROMADB_URL:=}
kvstore:
type: postgres
host: ${env.POSTGRES_HOST:=localhost}
port: ${env.POSTGRES_PORT:=5432}
db: ${env.POSTGRES_DB:=llamastack}
user: ${env.POSTGRES_USER:=llamastack}
password: ${env.POSTGRES_PASSWORD:=llamastack}
persistence:
namespace: vector_io::chroma_remote
backend: kv_default
files:
- provider_id: meta-reference-files
provider_type: inline::localfs
config:
storage_dir: ${env.FILES_STORAGE_DIR:=~/.llama/distributions/starter/files}
metadata_store:
type: sqlite
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/starter}/files_metadata.db
table_name: files_metadata
backend: sql_default
safety:
- provider_id: llama-guard
provider_type: inline::llama-guard
@ -56,26 +51,15 @@ providers:
- provider_id: meta-reference
provider_type: inline::meta-reference
config:
persistence_store:
type: postgres
host: ${env.POSTGRES_HOST:=localhost}
port: ${env.POSTGRES_PORT:=5432}
db: ${env.POSTGRES_DB:=llamastack}
user: ${env.POSTGRES_USER:=llamastack}
password: ${env.POSTGRES_PASSWORD:=llamastack}
responses_store:
type: postgres
host: ${env.POSTGRES_HOST:=localhost}
port: ${env.POSTGRES_PORT:=5432}
db: ${env.POSTGRES_DB:=llamastack}
user: ${env.POSTGRES_USER:=llamastack}
password: ${env.POSTGRES_PASSWORD:=llamastack}
telemetry:
- provider_id: meta-reference
provider_type: inline::meta-reference
config:
service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
sinks: ${env.TELEMETRY_SINKS:=console}
persistence:
agent_state:
namespace: agents
backend: kv_default
responses:
table_name: responses
backend: sql_default
max_write_queue_size: 10000
num_writers: 4
tool_runtime:
- provider_id: brave-search
provider_type: remote::brave-search
@ -110,40 +94,54 @@ storage:
db: ${env.POSTGRES_DB:=llamastack}
user: ${env.POSTGRES_USER:=llamastack}
password: ${env.POSTGRES_PASSWORD:=llamastack}
references:
stores:
metadata:
backend: kv_default
namespace: registry
backend: kv_default
inference:
backend: sql_default
table_name: inference_store
models:
- metadata:
embedding_dimension: 768
model_id: nomic-embed-text-v1.5
provider_id: sentence-transformers
model_type: embedding
- metadata: {}
model_id: ${env.INFERENCE_MODEL}
provider_id: vllm-inference
model_type: llm
- metadata: {}
model_id: ${env.SAFETY_MODEL:=meta-llama/Llama-Guard-3-1B}
provider_id: vllm-safety
model_type: llm
shields:
- shield_id: ${env.SAFETY_MODEL:=meta-llama/Llama-Guard-3-1B}
vector_dbs: []
datasets: []
scoring_fns: []
benchmarks: []
tool_groups:
- toolgroup_id: builtin::websearch
provider_id: tavily-search
- toolgroup_id: builtin::rag
provider_id: rag-runtime
backend: sql_default
max_write_queue_size: 10000
num_writers: 4
conversations:
table_name: openai_conversations
backend: sql_default
prompts:
namespace: prompts
backend: kv_default
registered_resources:
models:
- metadata:
embedding_dimension: 768
model_id: nomic-embed-text-v1.5
provider_id: sentence-transformers
model_type: embedding
- metadata: {}
model_id: ${env.INFERENCE_MODEL}
provider_id: vllm-inference
model_type: llm
- metadata: {}
model_id: ${env.SAFETY_MODEL:=meta-llama/Llama-Guard-3-1B}
provider_id: vllm-safety
model_type: llm
shields:
- shield_id: ${env.SAFETY_MODEL:=meta-llama/Llama-Guard-3-1B}
vector_dbs: []
datasets: []
scoring_fns: []
benchmarks: []
tool_groups:
- toolgroup_id: builtin::websearch
provider_id: tavily-search
- toolgroup_id: builtin::rag
provider_id: rag-runtime
server:
port: 8321
auth:
provider_config:
type: github_token
vector_stores:
default_provider_id: chromadb
default_embedding_model:
provider_id: sentence-transformers
model_id: nomic-ai/nomic-embed-text-v1.5

2
docs/docs/distributions/k8s/ui-k8s.yaml.template
View file

@ -44,7 +44,7 @@ spec:
# Navigate to the UI directory
echo "Navigating to UI directory..."
cd /app/llama_stack/ui
cd /app/llama_stack_ui
# Check if package.json exists
if [ ! -f "package.json" ]; then

4
docs/docs/distributions/list_of_distributions.mdx
View file

@ -28,7 +28,7 @@ Llama Stack provides several pre-configured distributions to help you get starte
- Run locally with Ollama for development
```bash
docker pull llama-stack/distribution-starter
docker pull llamastack/distribution-starter
```
**Guides:** [Starter Distribution Guide](self_hosted_distro/starter)
@ -41,7 +41,7 @@ docker pull llama-stack/distribution-starter
- Need to run inference locally
```bash
docker pull llama-stack/distribution-meta-reference-gpu
docker pull llamastack/distribution-meta-reference-gpu
```
**Guides:** [Meta Reference GPU Guide](self_hosted_distro/meta-reference-gpu)

109
docs/docs/distributions/llama_stack_ui.mdx Normal file
View file

@ -0,0 +1,109 @@
---
title: Llama Stack UI
description: Web-based user interface for interacting with Llama Stack servers
sidebar_label: Llama Stack UI
sidebar_position: 8
---
# Llama Stack UI
The Llama Stack UI is a web-based interface for interacting with Llama Stack servers. Built with Next.js and React, it provides a visual way to work with agents, manage resources, and view logs.
## Features
- **Logs & Monitoring**: View chat completions, agent responses, and vector store activity
- **Vector Stores**: Create and manage vector databases for RAG (Retrieval-Augmented Generation) workflows
- **Prompt Management**: Create and manage reusable prompts
## Prerequisites
You need a running Llama Stack server. The UI is a client that connects to the Llama Stack backend.
If you don't have a Llama Stack server running yet, see the [Starting Llama Stack Server](../getting_started/starting_llama_stack_server.mdx) guide.
## Running the UI
### Option 1: Using npx (Recommended for Quick Start)
The fastest way to get started is using `npx`:
```bash
npx llama-stack-ui
```
This will start the UI server on `http://localhost:8322` (default port).
### Option 2: Using Docker
Run the UI in a container:
```bash
docker run -p 8322:8322 llamastack/ui
```
Access the UI at `http://localhost:8322`.
## Environment Variables
The UI can be configured using the following environment variables:
| Variable | Description | Default |
|----------|-------------|---------|
| `LLAMA_STACK_BACKEND_URL` | URL of your Llama Stack server | `http://localhost:8321` |
| `LLAMA_STACK_UI_PORT` | Port for the UI server | `8322` |
If the Llama Stack server is running with authentication enabled, you can configure the UI to use it by setting the following environment variables:
| Variable | Description | Default |
|----------|-------------|---------|
| `NEXTAUTH_URL` | NextAuth URL for authentication | `http://localhost:8322` |
| `GITHUB_CLIENT_ID` | GitHub OAuth client ID (optional, for authentication) | - |
| `GITHUB_CLIENT_SECRET` | GitHub OAuth client secret (optional, for authentication) | - |
### Setting Environment Variables
#### For npx:
```bash
LLAMA_STACK_BACKEND_URL=http://localhost:8321 \
LLAMA_STACK_UI_PORT=8080 \
npx llama-stack-ui
```
#### For Docker:
```bash
docker run -p 8080:8080 \
-e LLAMA_STACK_BACKEND_URL=http://localhost:8321 \
-e LLAMA_STACK_UI_PORT=8080 \
llamastack/ui
```
## Using the UI
### Managing Resources
- **Vector Stores**: Create vector databases for RAG workflows, view stored documents and embeddings
- **Prompts**: Create and manage reusable prompt templates
- **Chat Completions**: View history of chat interactions
- **Responses**: Browse detailed agent responses and tool calls
## Development
If you want to run the UI from source for development:
```bash
# From the project root
cd src/llama_stack_ui
# Install dependencies
npm install
# Set environment variables
export LLAMA_STACK_BACKEND_URL=http://localhost:8321
# Start the development server
npm run dev
```
The development server will start on `http://localhost:8322` with hot reloading enabled.

6
docs/docs/distributions/remote_hosted_distro/index.mdx
View file

@ -2,10 +2,10 @@
Remote-Hosted distributions are available endpoints serving Llama Stack API that you can directly connect to.
| Distribution | Endpoint | Inference | Agents | Memory | Safety | Telemetry |
| Distribution | Endpoint | Inference | Agents | Memory | Safety |
|-------------|----------|-----------|---------|---------|---------|------------|
| Together | [https://llama-stack.together.ai](https://llama-stack.together.ai) | remote::together | meta-reference | remote::weaviate | meta-reference | meta-reference |
| Fireworks | [https://llamastack-preview.fireworks.ai](https://llamastack-preview.fireworks.ai) | remote::fireworks | meta-reference | remote::weaviate | meta-reference | meta-reference |
| Together | [https://llama-stack.together.ai](https://llama-stack.together.ai) | remote::together | meta-reference | remote::weaviate | meta-reference |
| Fireworks | [https://llamastack-preview.fireworks.ai](https://llamastack-preview.fireworks.ai) | remote::fireworks | meta-reference | remote::weaviate | meta-reference |
## Connecting to Remote-Hosted Distributions

143
docs/docs/distributions/remote_hosted_distro/oci.md Normal file
View file

@ -0,0 +1,143 @@
---
orphan: true
---
<!-- This file was auto-generated by distro_codegen.py, please edit source -->
# OCI Distribution
The `llamastack/distribution-oci` distribution consists of the following provider configurations.
| API | Provider(s) |
|-----|-------------|
| agents | `inline::meta-reference` |
| datasetio | `remote::huggingface`, `inline::localfs` |
| eval | `inline::meta-reference` |
| files | `inline::localfs` |
| inference | `remote::oci` |
| safety | `inline::llama-guard` |
| scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
| tool_runtime | `remote::brave-search`, `remote::tavily-search`, `inline::rag-runtime`, `remote::model-context-protocol` |
| vector_io | `inline::faiss`, `remote::chromadb`, `remote::pgvector` |
### Environment Variables
The following environment variables can be configured:
- `OCI_AUTH_TYPE`: OCI authentication type (instance_principal or config_file) (default: `instance_principal`)
- `OCI_REGION`: OCI region (e.g., us-ashburn-1, us-chicago-1, us-phoenix-1, eu-frankfurt-1) (default: ``)
- `OCI_COMPARTMENT_OCID`: OCI compartment ID for the Generative AI service (default: ``)
- `OCI_CONFIG_FILE_PATH`: OCI config file path (required if OCI_AUTH_TYPE is config_file) (default: `~/.oci/config`)
- `OCI_CLI_PROFILE`: OCI CLI profile name to use from config file (default: `DEFAULT`)
## Prerequisites
### Oracle Cloud Infrastructure Setup
Before using the OCI Generative AI distribution, ensure you have:
1. **Oracle Cloud Infrastructure Account**: Sign up at [Oracle Cloud Infrastructure](https://cloud.oracle.com/)
2. **Generative AI Service Access**: Enable the Generative AI service in your OCI tenancy
3. **Compartment**: Create or identify a compartment where you'll deploy Generative AI models
4. **Authentication**: Configure authentication using either:
- **Instance Principal** (recommended for cloud-hosted deployments)
- **API Key** (for on-premises or development environments)
### Authentication Methods
#### Instance Principal Authentication (Recommended)
Instance Principal authentication allows OCI resources to authenticate using the identity of the compute instance they're running on. This is the most secure method for production deployments.
Requirements:
- Instance must be running in an Oracle Cloud Infrastructure compartment
- Instance must have appropriate IAM policies to access Generative AI services
#### API Key Authentication
For development or on-premises deployments, follow [this doc](https://docs.oracle.com/en-us/iaas/Content/API/Concepts/apisigningkey.htm) to learn how to create your API signing key for your config file.
### Required IAM Policies
Ensure your OCI user or instance has the following policy statements:
```
Allow group <group_name> to use generative-ai-inference-endpoints in compartment <compartment_name>
Allow group <group_name> to manage generative-ai-inference-endpoints in compartment <compartment_name>
```
## Supported Services
### Inference: OCI Generative AI
Oracle Cloud Infrastructure Generative AI provides access to high-performance AI models through OCI's Platform-as-a-Service offering. The service supports:
- **Chat Completions**: Conversational AI with context awareness
- **Text Generation**: Complete prompts and generate text content
#### Available Models
Common OCI Generative AI models include access to Meta, Cohere, OpenAI, Grok, and more models.
### Safety: Llama Guard
For content safety and moderation, this distribution uses Meta's LlamaGuard model through the OCI Generative AI service to provide:
- Content filtering and moderation
- Policy compliance checking
- Harmful content detection
### Vector Storage: Multiple Options
The distribution supports several vector storage providers:
- **FAISS**: Local in-memory vector search
- **ChromaDB**: Distributed vector database
- **PGVector**: PostgreSQL with vector extensions
### Additional Services
- **Dataset I/O**: Local filesystem and Hugging Face integration
- **Tool Runtime**: Web search (Brave, Tavily) and RAG capabilities
- **Evaluation**: Meta reference evaluation framework
## Running Llama Stack with OCI
You can run the OCI distribution via Docker or local virtual environment.
### Via venv
If you've set up your local development environment, you can also build the image using your local virtual environment.
```bash
OCI_AUTH=$OCI_AUTH_TYPE OCI_REGION=$OCI_REGION OCI_COMPARTMENT_OCID=$OCI_COMPARTMENT_OCID llama stack run --port 8321 oci
```
### Configuration Examples
#### Using Instance Principal (Recommended for Production)
```bash
export OCI_AUTH_TYPE=instance_principal
export OCI_REGION=us-chicago-1
export OCI_COMPARTMENT_OCID=ocid1.compartment.oc1..<your-compartment-id>
```
#### Using API Key Authentication (Development)
```bash
export OCI_AUTH_TYPE=config_file
export OCI_CONFIG_FILE_PATH=~/.oci/config
export OCI_CLI_PROFILE=DEFAULT
export OCI_REGION=us-chicago-1
export OCI_COMPARTMENT_OCID=ocid1.compartment.oc1..your-compartment-id
```
## Regional Endpoints
OCI Generative AI is available in multiple regions. The service automatically routes to the appropriate regional endpoint based on your configuration. For a full list of regional model availability, visit:
https://docs.oracle.com/en-us/iaas/Content/generative-ai/overview.htm#regions
## Troubleshooting
### Common Issues
1. **Authentication Errors**: Verify your OCI credentials and IAM policies
2. **Model Not Found**: Ensure the model OCID is correct and the model is available in your region
3. **Permission Denied**: Check compartment permissions and Generative AI service access
4. **Region Unavailable**: Verify the specified region supports Generative AI services
### Getting Help
For additional support:
- [OCI Generative AI Documentation](https://docs.oracle.com/en-us/iaas/Content/generative-ai/home.htm)
- [Llama Stack Issues](https://github.com/meta-llama/llama-stack/issues)

1
docs/docs/distributions/remote_hosted_distro/watsonx.md
View file

@ -21,7 +21,6 @@ The `llamastack/distribution-watsonx` distribution consists of the following pro
| inference | `remote::watsonx`, `inline::sentence-transformers` |
| safety | `inline::llama-guard` |
| scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
| telemetry | `inline::meta-reference` |
| tool_runtime | `remote::brave-search`, `remote::tavily-search`, `inline::rag-runtime`, `remote::model-context-protocol` |
| vector_io | `inline::faiss` |

6
docs/docs/distributions/self_hosted_distro/dell-tgi.md
View file

@ -13,9 +13,9 @@ self
The `llamastack/distribution-tgi` distribution consists of the following provider configurations.
| **API** | **Inference** | **Agents** | **Memory** | **Safety** | **Telemetry** |
|----------------- |--------------- |---------------- |-------------------------------------------------- |---------------- |---------------- |
| **Provider(s)** | remote::tgi | meta-reference | meta-reference, remote::pgvector, remote::chroma | meta-reference | meta-reference |
| **API** | **Inference** | **Agents** | **Memory** | **Safety** |
|----------------- |--------------- |---------------- |-------------------------------------------------- |---------------- |
| **Provider(s)** | remote::tgi | meta-reference | meta-reference, remote::pgvector, remote::chroma | meta-reference |
The only difference vs. the `tgi` distribution is that it runs the Dell-TGI server for inference.

1
docs/docs/distributions/self_hosted_distro/dell.md
View file

@ -22,7 +22,6 @@ The `llamastack/distribution-dell` distribution consists of the following provid
| inference | `remote::tgi`, `inline::sentence-transformers` |
| safety | `inline::llama-guard` |
| scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
| telemetry | `inline::meta-reference` |
| tool_runtime | `remote::brave-search`, `remote::tavily-search`, `inline::rag-runtime` |
| vector_io | `inline::faiss`, `remote::chromadb`, `remote::pgvector` |

27
docs/docs/distributions/self_hosted_distro/meta-reference-gpu.md
View file

@ -79,6 +79,33 @@ docker run \
--port $LLAMA_STACK_PORT
```
### Via Docker with Custom Run Configuration
You can also run the Docker container with a custom run configuration file by mounting it into the container:
```bash
# Set the path to your custom run.yaml file
CUSTOM_RUN_CONFIG=/path/to/your/custom-run.yaml
LLAMA_STACK_PORT=8321
docker run \
-it \
--pull always \
--gpu all \
-p $LLAMA_STACK_PORT:$LLAMA_STACK_PORT \
-v ~/.llama:/root/.llama \
-v $CUSTOM_RUN_CONFIG:/app/custom-run.yaml \
-e RUN_CONFIG_PATH=/app/custom-run.yaml \
llamastack/distribution-meta-reference-gpu \
--port $LLAMA_STACK_PORT
```
**Note**: The run configuration must be mounted into the container before it can be used. The `-v` flag mounts your local file into the container, and the `RUN_CONFIG_PATH` environment variable tells the entrypoint script which configuration to use.
Available run configurations for this distribution:
- `run.yaml`
- `run-with-safety.yaml`
### Via venv
Make sure you have the Llama Stack CLI available.

30
docs/docs/distributions/self_hosted_distro/nvidia.md
View file

@ -127,13 +127,39 @@ docker run \
-it \
--pull always \
-p $LLAMA_STACK_PORT:$LLAMA_STACK_PORT \
-v ./run.yaml:/root/my-run.yaml \
-v ~/.llama:/root/.llama \
-e NVIDIA_API_KEY=$NVIDIA_API_KEY \
llamastack/distribution-nvidia \
--config /root/my-run.yaml \
--port $LLAMA_STACK_PORT
```
### Via Docker with Custom Run Configuration
You can also run the Docker container with a custom run configuration file by mounting it into the container:
```bash
# Set the path to your custom run.yaml file
CUSTOM_RUN_CONFIG=/path/to/your/custom-run.yaml
LLAMA_STACK_PORT=8321
docker run \
-it \
--pull always \
-p $LLAMA_STACK_PORT:$LLAMA_STACK_PORT \
-v ~/.llama:/root/.llama \
-v $CUSTOM_RUN_CONFIG:/app/custom-run.yaml \
-e RUN_CONFIG_PATH=/app/custom-run.yaml \
-e NVIDIA_API_KEY=$NVIDIA_API_KEY \
llamastack/distribution-nvidia \
--port $LLAMA_STACK_PORT
```
**Note**: The run configuration must be mounted into the container before it can be used. The `-v` flag mounts your local file into the container, and the `RUN_CONFIG_PATH` environment variable tells the entrypoint script which configuration to use.
Available run configurations for this distribution:
- `run.yaml`
- `run-with-safety.yaml`
### Via venv
If you've set up your local development environment, you can also install the distribution dependencies using your local virtual environment.

1
docs/docs/distributions/self_hosted_distro/passthrough.md
View file

@ -21,7 +21,6 @@ The `llamastack/distribution-passthrough` distribution consists of the following
| inference | `remote::passthrough`, `inline::sentence-transformers` |
| safety | `inline::llama-guard` |
| scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
| telemetry | `inline::meta-reference` |
| tool_runtime | `remote::brave-search`, `remote::tavily-search`, `remote::wolfram-alpha`, `inline::rag-runtime`, `remote::model-context-protocol` |
| vector_io | `inline::faiss`, `remote::chromadb`, `remote::pgvector` |

48
docs/docs/distributions/self_hosted_distro/starter.md
View file

@ -26,7 +26,6 @@ The starter distribution consists of the following provider configurations:
| inference | `remote::openai`, `remote::fireworks`, `remote::together`, `remote::ollama`, `remote::anthropic`, `remote::gemini`, `remote::groq`, `remote::sambanova`, `remote::vllm`, `remote::tgi`, `remote::cerebras`, `remote::llama-openai-compat`, `remote::nvidia`, `remote::hf::serverless`, `remote::hf::endpoint`, `inline::sentence-transformers` |
| safety | `inline::llama-guard` |
| scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
| telemetry | `inline::meta-reference` |
| tool_runtime | `remote::brave-search`, `remote::tavily-search`, `inline::rag-runtime`, `remote::model-context-protocol` |
| vector_io | `inline::faiss`, `inline::sqlite-vec`, `inline::milvus`, `remote::chromadb`, `remote::pgvector` |
@ -117,10 +116,6 @@ The following environment variables can be configured:
- `BRAVE_SEARCH_API_KEY`: Brave Search API key
- `TAVILY_SEARCH_API_KEY`: Tavily Search API key
### Telemetry Configuration
- `OTEL_SERVICE_NAME`: OpenTelemetry service name
- `TELEMETRY_SINKS`: Telemetry sinks (default: `[]`)
## Enabling Providers
You can enable specific providers by setting appropriate environment variables. For example,
@ -164,7 +159,41 @@ docker run \
--port $LLAMA_STACK_PORT
```
### Via venv
The container will run the distribution with a SQLite store by default. This store is used for the following components:
- Metadata store: store metadata about the models, providers, etc.
- Inference store: collect of responses from the inference provider
- Agents store: store agent configurations (sessions, turns, etc.)
- Agents Responses store: store responses from the agents
However, you can use PostgreSQL instead by running the `starter::run-with-postgres-store.yaml` configuration:
```bash
docker run \
-it \
--pull always \
-p $LLAMA_STACK_PORT:$LLAMA_STACK_PORT \
-e OPENAI_API_KEY=your_openai_key \
-e FIREWORKS_API_KEY=your_fireworks_key \
-e TOGETHER_API_KEY=your_together_key \
-e POSTGRES_HOST=your_postgres_host \
-e POSTGRES_PORT=your_postgres_port \
-e POSTGRES_DB=your_postgres_db \
-e POSTGRES_USER=your_postgres_user \
-e POSTGRES_PASSWORD=your_postgres_password \
llamastack/distribution-starter \
starter::run-with-postgres-store.yaml
```
Postgres environment variables:
- `POSTGRES_HOST`: Postgres host (default: `localhost`)
- `POSTGRES_PORT`: Postgres port (default: `5432`)
- `POSTGRES_DB`: Postgres database name (default: `llamastack`)
- `POSTGRES_USER`: Postgres username (default: `llamastack`)
- `POSTGRES_PASSWORD`: Postgres password (default: `llamastack`)
### Via Conda or venv
Ensure you have configured the starter distribution using the environment variables explained above.
@ -172,8 +201,11 @@ Ensure you have configured the starter distribution using the environment variab
# Install dependencies for the starter distribution
uv run --with llama-stack llama stack list-deps starter | xargs -L1 uv pip install
# Run the server
# Run the server (with SQLite - default)
uv run --with llama-stack llama stack run starter
# Or run with PostgreSQL
uv run --with llama-stack llama stack run starter::run-with-postgres-store.yaml
```
## Example Usage
@ -229,7 +261,7 @@ The starter distribution uses SQLite for local storage of various components:
2. **Flexible Configuration**: Easy to enable/disable providers based on your needs
3. **No Local GPU Required**: Most providers are cloud-based, making it accessible to developers without high-end hardware
4. **Easy Migration**: Start with hosted providers and gradually move to local ones as needed
5. **Production Ready**: Includes safety, evaluation, and telemetry components
5. **Production Ready**: Includes safety and evaluation
6. **Tool Integration**: Comes with web search, RAG, and model context protocol tools
The starter distribution is ideal for developers who want to experiment with different AI providers, build prototypes quickly, or create applications that can work with multiple AI backends.

2
docs/docs/distributions/starting_llama_stack_server.mdx
View file

@ -27,7 +27,7 @@ If you have built a container image and want to deploy it in a Kubernetes cluste
Control log output via environment variables before starting the server.
- `LLAMA_STACK_LOGGING` sets per-component levels, e.g. `LLAMA_STACK_LOGGING=server=debug;core=info`.
- `LLAMA_STACK_LOGGING` sets per-component levels, e.g. `LLAMA_STACK_LOGGING=server=debug,core=info`.
- Supported categories: `all`, `core`, `server`, `router`, `inference`, `agents`, `safety`, `eval`, `tools`, `client`.
- Levels: `debug`, `info`, `warning`, `error`, `critical` (default is `info`). Use `all=<level>` to apply globally.
- `LLAMA_STACK_LOG_FILE=/path/to/log` mirrors logs to a file while still printing to stdout.

33
docs/docs/getting_started/detailed_tutorial.mdx
View file

@ -144,7 +144,7 @@ source .venv/bin/activate
```bash
uv venv client --python 3.12
source client/bin/activate
pip install llama-stack-client
uv pip install llama-stack-client
```
</TabItem>
</Tabs>
@ -239,8 +239,13 @@ client = LlamaStackClient(base_url="http://localhost:8321")
models = client.models.list()
# Select the first LLM
llm = next(m for m in models if m.model_type == "llm" and m.provider_id == "ollama")
model_id = llm.identifier
llm = next(
m for m in models
if m.custom_metadata
and m.custom_metadata.get("model_type") == "llm"
and m.custom_metadata.get("provider_id") == "ollama"
)
model_id = llm.id
print("Model:", model_id)
@ -279,8 +284,13 @@ import uuid
client = LlamaStackClient(base_url=f"http://localhost:8321")
models = client.models.list()
llm = next(m for m in models if m.model_type == "llm" and m.provider_id == "ollama")
model_id = llm.identifier
llm = next(
m for m in models
if m.custom_metadata
and m.custom_metadata.get("model_type") == "llm"
and m.custom_metadata.get("provider_id") == "ollama"
)
model_id = llm.id
agent = Agent(client, model=model_id, instructions="You are a helpful assistant.")
@ -450,8 +460,11 @@ import uuid
client = LlamaStackClient(base_url="http://localhost:8321")
# Create a vector database instance
embed_lm = next(m for m in client.models.list() if m.model_type == "embedding")
embedding_model = embed_lm.identifier
embed_lm = next(
m for m in client.models.list()
if m.custom_metadata and m.custom_metadata.get("model_type") == "embedding"
)
embedding_model = embed_lm.id
vector_db_id = f"v{uuid.uuid4().hex}"
# The VectorDB API is deprecated; the server now returns its own authoritative ID.
# We capture the correct ID from the response's .identifier attribute.
@ -489,9 +502,11 @@ client.tool_runtime.rag_tool.insert(
llm = next(
m
for m in client.models.list()
if m.model_type == "llm" and m.provider_id == "ollama"
if m.custom_metadata
and m.custom_metadata.get("model_type") == "llm"
and m.custom_metadata.get("provider_id") == "ollama"
)
model = llm.identifier
model = llm.id
# Create the RAG agent
rag_agent = Agent(

24
docs/docs/getting_started/quickstart.mdx
View file

@ -24,6 +24,9 @@ ollama run llama3.2:3b --keepalive 60m
#### Step 2: Run the Llama Stack server
```python file=./demo_script.py title="demo_script.py"
```
We will use `uv` to install dependencies and run the Llama Stack server.
```bash
# Install dependencies for the starter distribution
@ -35,27 +38,6 @@ OLLAMA_URL=http://localhost:11434 uv run --with llama-stack llama stack run star
#### Step 3: Run the demo
Now open up a new terminal and copy the following script into a file named `demo_script.py`.
```python
import io, requests
from openai import OpenAI
url="https://www.paulgraham.com/greatwork.html"
client = OpenAI(base_url="http://localhost:8321/v1/", api_key="none")
vs = client.vector_stores.create()
response = requests.get(url)
pseudo_file = io.BytesIO(str(response.content).encode('utf-8'))
uploaded_file = client.files.create(file=(url, pseudo_file, "text/html"), purpose="assistants")
client.vector_stores.files.create(vector_store_id=vs.id, file_id=uploaded_file.id)
resp = client.responses.create(
model="openai/gpt-4o",
input="How do you do great work? Use the existing knowledge_search tool.",
tools=[{"type": "file_search", "vector_store_ids": [vs.id]}],
include=["file_search_call.results"],
)
We will use `uv` to run the script
```
uv run --with llama-stack-client,fire,requests demo_script.py

2
docs/docs/index.mdx
View file

@ -29,7 +29,7 @@ Llama Stack is now available! See the [release notes](https://github.com/llamast
Llama Stack defines and standardizes the core building blocks needed to bring generative AI applications to market. It provides a unified set of APIs with implementations from leading service providers, enabling seamless transitions between development and production environments. More specifically, it provides:
- **Unified API layer** for Inference, RAG, Agents, Tools, Safety, Evals, and Telemetry.
- **Unified API layer** for Inference, RAG, Agents, Tools, Safety, Evals.
- **Plugin architecture** to support the rich ecosystem of implementations of the different APIs in different environments like local development, on-premises, cloud, and mobile.
- **Prepackaged verified distributions** which offer a one-stop solution for developers to get started quickly and reliably in any environment
- **Multiple developer interfaces** like CLI and SDKs for Python, Node, iOS, and Android

7
docs/docs/providers/agents/index.mdx
View file

@ -1,7 +1,8 @@
---
description: "Agents
description: |
Agents
APIs for creating and interacting with agentic systems."
APIs for creating and interacting with agentic systems.
sidebar_label: Agents
title: Agents
---
@ -12,6 +13,6 @@ title: Agents
Agents
APIs for creating and interacting with agentic systems.
APIs for creating and interacting with agentic systems.
This section contains documentation for all available providers for the **agents** API.

2
docs/docs/providers/agents/inline_meta-reference.mdx
View file

@ -14,7 +14,7 @@ Meta's reference implementation of an agent system that can use tools, access ve
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `persistence` | `<class 'inline.agents.meta_reference.config.AgentPersistenceConfig'>` | No | | |
| `persistence` | `AgentPersistenceConfig` | No | | |
## Sample Configuration

27
docs/docs/providers/batches/index.mdx
View file

@ -1,14 +1,15 @@
---
description: "The Batches API enables efficient processing of multiple requests in a single operation,
particularly useful for processing large datasets, batch evaluation workflows, and
cost-effective inference at scale.
description: |
The Batches API enables efficient processing of multiple requests in a single operation,
particularly useful for processing large datasets, batch evaluation workflows, and
cost-effective inference at scale.
The API is designed to allow use of openai client libraries for seamless integration.
The API is designed to allow use of openai client libraries for seamless integration.
This API provides the following extensions:
- idempotent batch creation
This API provides the following extensions:
- idempotent batch creation
Note: This API is currently under active development and may undergo changes."
Note: This API is currently under active development and may undergo changes.
sidebar_label: Batches
title: Batches
---
@ -18,14 +19,14 @@ title: Batches
## Overview
The Batches API enables efficient processing of multiple requests in a single operation,
particularly useful for processing large datasets, batch evaluation workflows, and
cost-effective inference at scale.
particularly useful for processing large datasets, batch evaluation workflows, and
cost-effective inference at scale.
The API is designed to allow use of openai client libraries for seamless integration.
The API is designed to allow use of openai client libraries for seamless integration.
This API provides the following extensions:
- idempotent batch creation
This API provides the following extensions:
- idempotent batch creation
Note: This API is currently under active development and may undergo changes.
Note: This API is currently under active development and may undergo changes.
This section contains documentation for all available providers for the **batches** API.

6
docs/docs/providers/batches/inline_reference.mdx
View file

@ -14,9 +14,9 @@ Reference implementation of batches API with KVStore persistence.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `kvstore` | `<class 'llama_stack.core.storage.datatypes.KVStoreReference'>` | No | | Configuration for the key-value store backend. |
| `max_concurrent_batches` | `<class 'int'>` | No | 1 | Maximum number of concurrent batches to process simultaneously. |
| `max_concurrent_requests_per_batch` | `<class 'int'>` | No | 10 | Maximum number of concurrent requests to process per batch. |
| `kvstore` | `KVStoreReference` | No | | Configuration for the key-value store backend. |
| `max_concurrent_batches` | `int` | No | 1 | Maximum number of concurrent batches to process simultaneously. |
| `max_concurrent_requests_per_batch` | `int` | No | 10 | Maximum number of concurrent requests to process per batch. |
## Sample Configuration

2
docs/docs/providers/datasetio/inline_localfs.mdx
View file

@ -14,7 +14,7 @@ Local filesystem-based dataset I/O provider for reading and writing datasets to
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `kvstore` | `<class 'llama_stack.core.storage.datatypes.KVStoreReference'>` | No | | |
| `kvstore` | `KVStoreReference` | No | | |
## Sample Configuration

2
docs/docs/providers/datasetio/remote_huggingface.mdx
View file

@ -14,7 +14,7 @@ HuggingFace datasets provider for accessing and managing datasets from the Huggi
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `kvstore` | `<class 'llama_stack.core.storage.datatypes.KVStoreReference'>` | No | | |
| `kvstore` | `KVStoreReference` | No | | |
## Sample Configuration

2
docs/docs/providers/datasetio/remote_nvidia.mdx
View file

@ -17,7 +17,7 @@ NVIDIA's dataset I/O provider for accessing datasets from NVIDIA's data platform
| `api_key` | `str \| None` | No | | The NVIDIA API key. |
| `dataset_namespace` | `str \| None` | No | default | The NVIDIA dataset namespace. |
| `project_id` | `str \| None` | No | test-project | The NVIDIA project ID. |
| `datasets_url` | `<class 'str'>` | No | http://nemo.test | Base URL for the NeMo Dataset API |
| `datasets_url` | `str` | No | http://nemo.test | Base URL for the NeMo Dataset API |
## Sample Configuration

7
docs/docs/providers/eval/index.mdx
View file

@ -1,7 +1,8 @@
---
description: "Evaluations
description: |
Evaluations
Llama Stack Evaluation API for running evaluations on model and agent candidates."
Llama Stack Evaluation API for running evaluations on model and agent candidates.
sidebar_label: Eval
title: Eval
---
@ -12,6 +13,6 @@ title: Eval
Evaluations
Llama Stack Evaluation API for running evaluations on model and agent candidates.
Llama Stack Evaluation API for running evaluations on model and agent candidates.
This section contains documentation for all available providers for the **eval** API.

2
docs/docs/providers/eval/inline_meta-reference.mdx
View file

@ -14,7 +14,7 @@ Meta's reference implementation of evaluation tasks with support for multiple la
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `kvstore` | `<class 'llama_stack.core.storage.datatypes.KVStoreReference'>` | No | | |
| `kvstore` | `KVStoreReference` | No | | |
## Sample Configuration

2
docs/docs/providers/eval/remote_nvidia.mdx
View file

@ -14,7 +14,7 @@ NVIDIA's evaluation provider for running evaluation tasks on NVIDIA's platform.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `evaluator_url` | `<class 'str'>` | No | http://0.0.0.0:7331 | The url for accessing the evaluator service |
| `evaluator_url` | `str` | No | http://0.0.0.0:7331 | The url for accessing the evaluator service |
## Sample Configuration

2
docs/docs/providers/external/external-providers-guide.mdx vendored
View file

@ -80,7 +80,7 @@ container_image: custom-vector-store:latest # optional
All providers must contain a `get_provider_spec` function in their `provider` module. This is a standardized structure that Llama Stack expects and is necessary for getting things such as the config class. The `get_provider_spec` method returns a structure identical to the `adapter`. An example function may look like:
```python
from llama_stack.providers.datatypes import (
from llama_stack_api.providers.datatypes import (
ProviderSpec,
Api,
RemoteProviderSpec,

290
docs/docs/providers/files/files.mdx Normal file
View file

@ -0,0 +1,290 @@
---
sidebar_label: Files
title: Files
---
## Overview
The Files API provides file management capabilities for Llama Stack. It allows you to upload, store, retrieve, and manage files that can be used across various endpoints in your application.
## Features
- **File Upload**: Upload files with metadata and purpose classification
- **File Management**: List, retrieve, and delete files
- **Content Retrieval**: Access raw file content for processing
- **API Compatibility**: Full compatibility with OpenAI Files API endpoints
- **Flexible Storage**: Support for local filesystem and cloud storage backends
## API Endpoints
### Upload File
**POST** `/v1/openai/v1/files`
Upload a file that can be used across various endpoints.
**Request Body:**
- `file`: The file object to be uploaded (multipart form data)
- `purpose`: The intended purpose of the uploaded file
**Supported Purposes:**
- `batch`: Files for batch operations
**Response:**
```json
{
"id": "file-abc123",
"object": "file",
"bytes": 140,
"created_at": 1613779121,
"filename": "mydata.jsonl",
"purpose": "batch"
}
```
**Example:**
```python
import requests
with open("data.jsonl", "rb") as f:
files = {"file": f}
data = {"purpose": "batch"}
response = requests.post(
"http://localhost:8000/v1/openai/v1/files", files=files, data=data
)
file_info = response.json()
```
### List Files
**GET** `/v1/openai/v1/files`
Returns a list of files that belong to the user's organization.
**Query Parameters:**
- `after` (optional): A cursor for pagination
- `limit` (optional): Limit on number of objects (1-10,000, default: 10,000)
- `order` (optional): Sort order by created_at timestamp (`asc` or `desc`, default: `desc`)
- `purpose` (optional): Filter files by purpose
**Response:**
```json
{
"object": "list",
"data": [
{
"id": "file-abc123",
"object": "file",
"bytes": 140,
"created_at": 1613779121,
"filename": "mydata.jsonl",
"purpose": "fine-tune"
}
],
"has_more": false
}
```
**Example:**
```python
import requests
# List all files
response = requests.get("http://localhost:8000/v1/openai/v1/files")
files = response.json()
# List files with pagination
response = requests.get(
"http://localhost:8000/v1/openAi/v1/files",
params={"limit": 10, "after": "file-abc123"},
)
files = response.json()
# Filter by purpose
response = requests.get(
"http://localhost:8000/v1/openAi/v1/files", params={"purpose": "fine-tune"}
)
files = response.json()
```
### Retrieve File
**GET** `/v1/openAi/v1/files/{file_id}`
Returns information about a specific file.
**Path Parameters:**
- `file_id`: The ID of the file to retrieve
**Response:**
```json
{
"id": "file-abc123",
"object": "file",
"bytes": 140,
"created_at": 1613779121,
"filename": "mydata.jsonl",
"purpose": "fine-tune"
}
```
**Example:**
```python
import requests
file_id = "file-abc123"
response = requests.get(f"http://localhost:8000/v1/openAi/v1/files/{file_id}")
file_info = response.json()
```
### Delete File
**DELETE** `/v1/openAi/v1/files/{file_id}`
Delete a file.
**Path Parameters:**
- `file_id`: The ID of the file to delete
**Response:**
```json
{
"id": "file-abc123",
"object": "file",
"deleted": true
}
```
**Example:**
```python
import requests
file_id = "file-abc123"
response = requests.delete(f"http://localhost:8000/v1/openAi/v1/files/{file_id}")
result = response.json()
```
### Retrieve File Content
**GET** `/v1/openAi/v1/files/{file_id}/content`
Returns the raw file content as a binary response.
**Path Parameters:**
- `file_id`: The ID of the file to retrieve content from
**Response:**
Binary file content with appropriate headers:
- `Content-Type`: `application/octet-stream`
- `Content-Disposition`: `attachment; filename="filename"`
**Example:**
```python
import requests
file_id = "file-abc123"
response = requests.get(f"http://localhost:8000/v1/openAi/v1/files/{file_id}/content")
# Save content to file
with open("downloaded_file.jsonl", "wb") as f:
f.write(response.content)
# Or process content directly
content = response.content
```
## Vector Store Integration
The Files API integrates with Vector Stores to enable document processing and search. For detailed information about this integration, see [File Operations and Vector Store Integration](../concepts/file_operations_vector_stores.md).
### Vector Store File Operations
**List Vector Store Files:**
- **GET** `/v1/openAi/v1/vector_stores/{vector_store_id}/files`
**Retrieve Vector Store File Content:**
- **GET** `/v1/openAi/v1/vector_stores/{vector_store_id}/files/{file_id}/content`
**Attach File to Vector Store:**
- **POST** `/v1/openAi/v1/vector_stores/{vector_store_id}/files`
## Error Handling
The Files API returns standard HTTP status codes and error responses:
- `400 Bad Request`: Invalid request parameters
- `404 Not Found`: File not found
- `429 Too Many Requests`: Rate limit exceeded
- `500 Internal Server Error`: Server error
**Error Response Format:**
```json
{
"error": {
"message": "Error description",
"type": "invalid_request_error",
"code": "file_not_found"
}
}
```
## Rate Limits
The Files API implements rate limiting to ensure fair usage:
- File uploads: 100 files per minute
- File retrievals: 1000 requests per minute
- File deletions: 100 requests per minute
## Best Practices
1. **File Organization**: Use descriptive filenames and appropriate purpose classifications
2. **Batch Operations**: For multiple files, consider using batch endpoints when available
3. **Error Handling**: Always check response status codes and handle errors gracefully
4. **Content Types**: Ensure files are uploaded with appropriate content types
5. **Cleanup**: Regularly delete unused files to manage storage costs
## Integration Examples
### With Python Client
```python
from llama_stack import LlamaStackClient
client = LlamaStackClient("http://localhost:8000")
# Upload a file
with open("data.jsonl", "rb") as f:
file_info = await client.files.upload(file=f, purpose="fine-tune")
# List files
files = await client.files.list(purpose="fine-tune")
# Retrieve file content
content = await client.files.retrieve_content(file_info.id)
```
### With cURL
```bash
# Upload file
curl -X POST http://localhost:8000/v1/openAi/v1/files \
-F "file=@data.jsonl" \
-F "purpose=fine-tune"
# List files
curl http://localhost:8000/v1/openAi/v1/files
# Download file content
curl http://localhost:8000/v1/openAi/v1/files/file-abc123/content \
-o downloaded_file.jsonl
```
## Provider Support
The Files API supports multiple storage backends:
- **Local Filesystem**: Store files on local disk (inline provider)
- **S3**: Store files in AWS S3 or S3-compatible services (remote provider)
- **Custom Backends**: Extensible architecture for custom storage providers
See the [Files Providers](index.md) documentation for detailed configuration options.

7
docs/docs/providers/files/index.mdx
View file

@ -1,7 +1,8 @@
---
description: "Files
description: |
Files
This API is used to upload documents that can be used with other Llama Stack APIs."
This API is used to upload documents that can be used with other Llama Stack APIs.
sidebar_label: Files
title: Files
---
@ -12,6 +13,6 @@ title: Files
Files
This API is used to upload documents that can be used with other Llama Stack APIs.
This API is used to upload documents that can be used with other Llama Stack APIs.
This section contains documentation for all available providers for the **files** API.

6
docs/docs/providers/files/inline_localfs.mdx
View file

@ -14,9 +14,9 @@ Local filesystem-based file storage provider for managing files and documents lo
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `storage_dir` | `<class 'str'>` | No | | Directory to store uploaded files |
| `metadata_store` | `<class 'llama_stack.core.storage.datatypes.SqlStoreReference'>` | No | | SQL store configuration for file metadata |
| `ttl_secs` | `<class 'int'>` | No | 31536000 | |
| `storage_dir` | `str` | No | | Directory to store uploaded files |
| `metadata_store` | `SqlStoreReference` | No | | SQL store configuration for file metadata |
| `ttl_secs` | `int` | No | 31536000 | |
## Sample Configuration

80
docs/docs/providers/files/openai_file_operations_quick_reference.md Normal file
View file

@ -0,0 +1,80 @@
# File Operations Quick Reference
## Overview
As of release 0.2.14, Llama Stack provides comprehensive file operations and Vector Store API integration, following the [OpenAI Vector Store Files API specification](https://platform.openai.com/docs/api-reference/vector-stores-files).
> **Note**: For detailed overview and implementation details, see [Overview](../openai_file_operations_support.md#overview) in the full documentation.
## Supported Providers
> **Note**: For complete provider details and features, see [Supported Providers](../openai_file_operations_support.md#supported-providers) in the full documentation.
**Inline Providers**: FAISS, SQLite-vec, Milvus
**Remote Providers**: ChromaDB, Qdrant, Weaviate, PGVector
## Quick Start
### 1. Upload File
```python
file_info = await client.files.upload(
file=open("document.pdf", "rb"), purpose="assistants"
)
```
### 2. Create Vector Store
```python
vector_store = client.vector_stores.create(name="my_docs")
```
### 3. Attach File
```python
await client.vector_stores.files.create(
vector_store_id=vector_store.id, file_id=file_info.id
)
```
### 4. Search
```python
results = await client.vector_stores.search(
vector_store_id=vector_store.id, query="What is the main topic?", max_num_results=5
)
```
## File Processing & Search
**Processing**: 800 tokens default chunk size, 400 token overlap
**Formats**: PDF, DOCX, TXT, Code files, etc.
**Search**: Vector similarity, Hybrid (SQLite-vec), Filtered with metadata
## Configuration
> **Note**: For detailed configuration examples and options, see [Configuration Examples](../openai_file_operations_support.md#configuration-examples) in the full documentation.
**Basic Setup**: Configure vector_io and files providers in your run.yaml
## Common Use Cases
- **RAG Systems**: Document Q&A with file uploads
- **Knowledge Bases**: Searchable document collections
- **Content Analysis**: Document similarity and clustering
- **Research Tools**: Literature review and analysis
## Performance Tips
> **Note**: For detailed performance optimization strategies, see [Performance Considerations](../openai_file_operations_support.md#performance-considerations) in the full documentation.
**Quick Tips**: Choose provider based on your needs (speed vs. storage vs. scalability)
## Troubleshooting
> **Note**: For comprehensive troubleshooting, see [Troubleshooting](../openai_file_operations_support.md#troubleshooting) in the full documentation.
**Quick Fixes**: Check file format compatibility, optimize chunk sizes, monitor storage
## Resources
- [Full Documentation](openai_file_operations_support.md)
- [Integration Guide](../concepts/file_operations_vector_stores.md)
- [Files API](files_api.md)
- [Provider Details](../vector_io/index.md)

291
docs/docs/providers/files/openai_file_operations_support.md Normal file
View file

@ -0,0 +1,291 @@
# File Operations Support in Vector Store Providers
## Overview
This document provides a comprehensive overview of file operations and Vector Store API support across all available vector store providers in Llama Stack. As of release 0.2.24, the following providers support full file operations integration.
## Supported Providers
### ✅ Full File Operations Support
The following providers support complete file operations integration, including file upload, automatic processing, and search:
#### Inline Providers (Single Node)
| Provider | File Operations | Key Features |
|----------|----------------|--------------|
| **FAISS** | ✅ Full Support | Fast in-memory search, GPU acceleration |
| **SQLite-vec** | ✅ Full Support | Hybrid search, disk-based storage |
| **Milvus** | ✅ Full Support | High-performance, scalable indexing |
#### Remote Providers (Hosted)
| Provider | File Operations | Key Features |
|----------|----------------|--------------|
| **ChromaDB** | ✅ Full Support | Metadata filtering, persistent storage |
| **Qdrant** | ✅ Full Support | Payload filtering, advanced search |
| **Weaviate** | ✅ Full Support | GraphQL interface, schema management |
| **Postgres (PGVector)** | ✅ Full Support | SQL integration, ACID compliance |
### 🔄 Partial Support
Some providers may support basic vector operations but lack full file operations integration:
| Provider | Status | Notes |
|----------|--------|-------|
| **Meta Reference** | 🔄 Basic | Core vector operations only |
## File Operations Features
All supported providers offer the following file operations capabilities:
### Core Functionality
- **File Upload & Processing**: Automatic document ingestion and chunking
- **Vector Storage**: Embedding generation and storage
- **Search & Retrieval**: Semantic search with metadata filtering
- **File Management**: List, retrieve, and manage files in vector stores
### Advanced Features
- **Automatic Chunking**: Configurable chunk sizes and overlap
- **Metadata Preservation**: File attributes and chunk metadata
- **Status Tracking**: Monitor file processing progress
- **Error Handling**: Comprehensive error reporting and recovery
## Implementation Details
### File Processing Pipeline
1. **Upload**: File uploaded via Files API
2. **Extraction**: Text content extracted from various formats
3. **Chunking**: Content split into optimal chunks (default: 800 tokens)
4. **Embedding**: Chunks converted to vector embeddings
5. **Storage**: Vectors stored with metadata in vector database
6. **Indexing**: Search index updated for fast retrieval
### Supported File Formats
- **Documents**: PDF, DOCX, DOC
- **Text**: TXT, MD, RST
- **Code**: Python, JavaScript, Java, C++, etc.
- **Data**: JSON, CSV, XML
- **Web**: HTML files
### Chunking Strategies
- **Default**: 800 tokens with 400 token overlap
- **Custom**: Configurable chunk sizes and overlap
- **Static**: Fixed-size chunks with overlap
## Provider-Specific Features
### FAISS
- **Storage**: In-memory with optional persistence
- **Performance**: Optimized for speed and GPU acceleration
- **Use Case**: High-performance, memory-constrained environments
### SQLite-vec
- **Storage**: Disk-based with SQLite backend
- **Search**: Hybrid vector + keyword search
- **Use Case**: Large document collections, frequent updates
### Milvus
- **Storage**: Scalable distributed storage
- **Indexing**: Multiple index types (IVF, HNSW)
- **Use Case**: Production deployments, large-scale applications
### ChromaDB
- **Storage**: Persistent storage with metadata
- **Filtering**: Advanced metadata filtering
- **Use Case**: Applications requiring rich metadata
### Qdrant
- **Storage**: High-performance vector database
- **Filtering**: Payload-based filtering
- **Use Case**: Real-time applications, complex queries
### Weaviate
- **Storage**: GraphQL-native vector database
- **Schema**: Flexible schema management
- **Use Case**: Applications requiring complex data relationships
### Postgres (PGVector)
- **Storage**: SQL database with vector extensions
- **Integration**: ACID compliance, existing SQL workflows
- **Use Case**: Applications requiring transactional guarantees
## Configuration Examples
### Basic Configuration
```yaml
vector_io:
- provider_id: faiss
provider_type: inline::faiss
config:
kvstore:
type: sqlite
db_path: ~/.llama/faiss_store.db
```
### With FileResponse Support
```yaml
vector_io:
- provider_id: faiss
provider_type: inline::faiss
config:
kvstore:
type: sqlite
db_path: ~/.llama/faiss_store.db
files:
- provider_id: local-files
provider_type: inline::localfs
config:
storage_dir: ~/.llama/files
metadata_store:
type: sqlite
db_path: ~/.llama/files_metadata.db
```
## Usage Examples
### Python Client
```python
from llama_stack import LlamaStackClient
client = LlamaStackClient("http://localhost:8000")
# Create vector store
vector_store = client.vector_stores.create(name="documents")
# Upload and process file
with open("document.pdf", "rb") as f:
file_info = await client.files.upload(file=f, purpose="assistants")
# Attach to vector store
await client.vector_stores.files.create(
vector_store_id=vector_store.id, file_id=file_info.id
)
# Search
results = await client.vector_stores.search(
vector_store_id=vector_store.id, query="What is the main topic?", max_num_results=5
)
```
### cURL Commands
```bash
# Upload file
curl -X POST http://localhost:8000/v1/openai/v1/files \
-F "file=@document.pdf" \
-F "purpose=assistants"
# Create vector store
curl -X POST http://localhost:8000/v1/openai/v1/vector_stores \
-H "Content-Type: application/json" \
-d '{"name": "documents"}'
# Attach file to vector store
curl -X POST http://localhost:8000/v1/openai/v1/vector_stores/{store_id}/files \
-H "Content-Type: application/json" \
-d '{"file_id": "file-abc123"}'
# Search vector store
curl -X POST http://localhost:8000/v1/openai/v1/vector_stores/{store_id}/search \
-H "Content-Type: application/json" \
-d '{"query": "What is the main topic?", "max_num_results": 5}'
```
## Performance Considerations
### Chunk Size Optimization
- **Small chunks (400-600 tokens)**: Better precision, more results
- **Large chunks (800-1200 tokens)**: Better context, fewer results
- **Overlap (50%)**: Maintains context between chunks
### Storage Efficiency
- **FAISS**: Fastest, but memory-limited
- **SQLite-vec**: Good balance of performance and storage
- **Milvus**: Scalable, production-ready
- **Remote providers**: Managed, but network-dependent
### Search Performance
- **Vector search**: Fastest for semantic queries
- **Hybrid search**: Best accuracy (SQLite-vec only)
- **Filtered search**: Fast with metadata constraints
## Troubleshooting
### Common Issues
1. **File Processing Failures**
- Check file format compatibility
- Verify file size limits
- Review error messages in file status
2. **Search Performance**
- Optimize chunk sizes for your use case
- Use filters to narrow search scope
- Monitor vector store metrics
3. **Storage Issues**
- Check available disk space
- Verify database permissions
- Monitor memory usage (for in-memory providers)
### Monitoring
```python
# Check file processing status
file_status = await client.vector_stores.files.retrieve(
vector_store_id=vector_store.id, file_id=file_info.id
)
if file_status.status == "failed":
print(f"Error: {file_status.last_error.message}")
# Monitor vector store health
health = await client.vector_stores.health(vector_store_id=vector_store.id)
print(f"Status: {health.status}")
```
## Best Practices
1. **File Organization**: Use descriptive names and organize by purpose
2. **Chunking Strategy**: Test different sizes for your specific use case
3. **Metadata**: Add relevant attributes for better filtering
4. **Monitoring**: Track processing status and search performance
5. **Cleanup**: Regularly remove unused files to manage storage
## Future Enhancements
Planned improvements for file operations support:
- **Batch Processing**: Process multiple files simultaneously
- **Advanced Chunking**: More sophisticated chunking algorithms
- **Custom Embeddings**: Support for custom embedding models
- **Real-time Updates**: Live file processing and indexing
- **Multi-format Support**: Enhanced file format support
## Support and Resources
- **Documentation**: [File Operations and Vector Store Integration](../../concepts/file_operations_vector_stores.mdx)
- **API Reference**: [Files API](files_api.md)
- **Provider Docs**: [Vector Store Providers](../vector_io/index.md)
- **Examples**: [Getting Started](../getting_started/index.md)
- **Community**: [GitHub Discussions](https://github.com/meta-llama/llama-stack/discussions)

27
docs/docs/providers/files/remote_openai.mdx Normal file
View file

@ -0,0 +1,27 @@
---
description: "OpenAI Files API provider for managing files through OpenAI's native file storage service."
sidebar_label: Remote - Openai
title: remote::openai
---
# remote::openai
## Description
OpenAI Files API provider for managing files through OpenAI's native file storage service.
## Configuration
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `api_key` | `str` | No | | OpenAI API key for authentication |
| `metadata_store` | `SqlStoreReference` | No | | SQL store configuration for file metadata |
## Sample Configuration
```yaml
api_key: ${env.OPENAI_API_KEY}
metadata_store:
table_name: openai_files_metadata
backend: sql_default
```

8
docs/docs/providers/files/remote_s3.mdx
View file

@ -14,13 +14,13 @@ AWS S3-based file storage provider for scalable cloud file management with metad
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `bucket_name` | `<class 'str'>` | No | | S3 bucket name to store files |
| `region` | `<class 'str'>` | No | us-east-1 | AWS region where the bucket is located |
| `bucket_name` | `str` | No | | S3 bucket name to store files |
| `region` | `str` | No | us-east-1 | AWS region where the bucket is located |
| `aws_access_key_id` | `str \| None` | No | | AWS access key ID (optional if using IAM roles) |
| `aws_secret_access_key` | `str \| None` | No | | AWS secret access key (optional if using IAM roles) |
| `endpoint_url` | `str \| None` | No | | Custom S3 endpoint URL (for MinIO, LocalStack, etc.) |
| `auto_create_bucket` | `<class 'bool'>` | No | False | Automatically create the S3 bucket if it doesn't exist |
| `metadata_store` | `<class 'llama_stack.core.storage.datatypes.SqlStoreReference'>` | No | | SQL store configuration for file metadata |
| `auto_create_bucket` | `bool` | No | False | Automatically create the S3 bucket if it doesn't exist |
| `metadata_store` | `SqlStoreReference` | No | | SQL store configuration for file metadata |
## Sample Configuration

16
docs/docs/providers/index.mdx
View file

@ -22,15 +22,25 @@ Importantly, Llama Stack always strives to provide at least one fully inline pro
## Provider Categories
- **[External Providers](external/index.mdx)** - Guide for building and using external providers
- **[OpenAI Compatibility](../api-openai/index.mdx)** - OpenAI API compatibility layer
- **[Inference](inference/index.mdx)** - LLM and embedding model providers
- **[Agents](agents/index.mdx)** - Agentic system providers
- **[DatasetIO](datasetio/index.mdx)** - Dataset and data loader providers
- **[Safety](safety/index.mdx)** - Content moderation and safety providers
- **[Telemetry](telemetry/index.mdx)** - Monitoring and observability providers
- **[Vector IO](vector_io/index.mdx)** - Vector database providers
- **[Tool Runtime](tool_runtime/index.mdx)** - Tool and protocol providers
- **[Files](files/index.mdx)** - File system and storage providers
## Other information about Providers
- **[OpenAI Compatibility](./openai.mdx)** - OpenAI API compatibility layer
## API Documentation
For comprehensive API documentation and reference:
- **[API Reference](../api/index.mdx)** - Complete API documentation
- **[Experimental APIs](../api-experimental/index.mdx)** - APIs in development
- **[Deprecated APIs](../api-deprecated/index.mdx)** - Legacy APIs being phased out
- **[OpenAI Compatibility](../api-openai/index.mdx)** - OpenAI API compatibility guide
## Additional Provider Information
- **[OpenAI Implementation Guide](./openai.mdx)** - Code examples and implementation details for OpenAI APIs
- **[OpenAI-Compatible Responses Limitations](./openai_responses_limitations.mdx)** - Known limitations of the Responses API in Llama Stack

21
docs/docs/providers/inference/index.mdx
View file

@ -1,11 +1,13 @@
---
description: "Inference
description: |
Inference
Llama Stack Inference API for generating completions, chat completions, and embeddings.
Llama Stack Inference API for generating completions, chat completions, and embeddings.
This API provides the raw interface to the underlying models. Two kinds of models are supported:
- LLM models: these models generate \"raw\" and \"chat\" (conversational) completions.
- Embedding models: these models generate embeddings to be used for semantic search."
This API provides the raw interface to the underlying models. Three kinds of models are supported:
- LLM models: these models generate "raw" and "chat" (conversational) completions.
- Embedding models: these models generate embeddings to be used for semantic search.
- Rerank models: these models reorder the documents based on their relevance to a query.
sidebar_label: Inference
title: Inference
---
@ -16,10 +18,11 @@ title: Inference
Inference
Llama Stack Inference API for generating completions, chat completions, and embeddings.
Llama Stack Inference API for generating completions, chat completions, and embeddings.
This API provides the raw interface to the underlying models. Two kinds of models are supported:
- LLM models: these models generate "raw" and "chat" (conversational) completions.
- Embedding models: these models generate embeddings to be used for semantic search.
This API provides the raw interface to the underlying models. Three kinds of models are supported:
- LLM models: these models generate "raw" and "chat" (conversational) completions.
- Embedding models: these models generate embeddings to be used for semantic search.
- Rerank models: these models reorder the documents based on their relevance to a query.
This section contains documentation for all available providers for the **inference** API.

8
docs/docs/providers/inference/inline_meta-reference.mdx
View file

@ -16,12 +16,12 @@ Meta's reference implementation of inference with support for various model form
|-------|------|----------|---------|-------------|
| `model` | `str \| None` | No | | |
| `torch_seed` | `int \| None` | No | | |
| `max_seq_len` | `<class 'int'>` | No | 4096 | |
| `max_batch_size` | `<class 'int'>` | No | 1 | |
| `max_seq_len` | `int` | No | 4096 | |
| `max_batch_size` | `int` | No | 1 | |
| `model_parallel_size` | `int \| None` | No | | |
| `create_distributed_process_group` | `<class 'bool'>` | No | True | |
| `create_distributed_process_group` | `bool` | No | True | |
| `checkpoint_dir` | `str \| None` | No | | |
| `quantization` | `Bf16QuantizationConfig \| Fp8QuantizationConfig \| Int4QuantizationConfig, annotation=NoneType, required=True, discriminator='type'` | No | | |
| `quantization` | `Bf16QuantizationConfig \| Fp8QuantizationConfig \| Int4QuantizationConfig \| None` | No | | |
## Sample Configuration

6
docs/docs/providers/inference/remote_anthropic.mdx
View file

@ -14,9 +14,9 @@ Anthropic inference provider for accessing Claude models and Anthropic's AI serv
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
## Sample Configuration

10
docs/docs/providers/inference/remote_azure.mdx
View file

@ -21,10 +21,10 @@ https://learn.microsoft.com/en-us/azure/ai-foundry/openai/overview
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `api_base` | `<class 'pydantic.networks.HttpUrl'>` | No | | Azure API base for Azure (e.g., https://your-resource-name.openai.azure.com) |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | | Azure API base for Azure (e.g., https://your-resource-name.openai.azure.com/openai/v1) |
| `api_version` | `str \| None` | No | | Azure API version for Azure (e.g., 2024-12-01-preview) |
| `api_type` | `str \| None` | No | azure | Azure API type for Azure (e.g., azure) |
@ -32,7 +32,7 @@ https://learn.microsoft.com/en-us/azure/ai-foundry/openai/overview
```yaml
api_key: ${env.AZURE_API_KEY:=}
api_base: ${env.AZURE_API_BASE:=}
base_url: ${env.AZURE_API_BASE:=}
api_version: ${env.AZURE_API_VERSION:=}
api_type: ${env.AZURE_API_TYPE:=}
```

23
docs/docs/providers/inference/remote_bedrock.mdx
View file

@ -1,5 +1,5 @@
---
description: "AWS Bedrock inference provider for accessing various AI models through AWS's managed service."
description: "AWS Bedrock inference provider using OpenAI compatible endpoint."
sidebar_label: Remote - Bedrock
title: remote::bedrock
---
@ -8,27 +8,20 @@ title: remote::bedrock
## Description
AWS Bedrock inference provider for accessing various AI models through AWS's managed service.
AWS Bedrock inference provider using OpenAI compatible endpoint.
## Configuration
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `aws_access_key_id` | `str \| None` | No | | The AWS access key to use. Default use environment variable: AWS_ACCESS_KEY_ID |
| `aws_secret_access_key` | `str \| None` | No | | The AWS secret access key to use. Default use environment variable: AWS_SECRET_ACCESS_KEY |
| `aws_session_token` | `str \| None` | No | | The AWS session token to use. Default use environment variable: AWS_SESSION_TOKEN |
| `region_name` | `str \| None` | No | | The default AWS Region to use, for example, us-west-1 or us-west-2.Default use environment variable: AWS_DEFAULT_REGION |
| `profile_name` | `str \| None` | No | | The profile name that contains credentials to use.Default use environment variable: AWS_PROFILE |
| `total_max_attempts` | `int \| None` | No | | An integer representing the maximum number of attempts that will be made for a single request, including the initial attempt. Default use environment variable: AWS_MAX_ATTEMPTS |
| `retry_mode` | `str \| None` | No | | A string representing the type of retries Boto3 will perform.Default use environment variable: AWS_RETRY_MODE |
| `connect_timeout` | `float \| None` | No | 60.0 | The time in seconds till a timeout exception is thrown when attempting to make a connection. The default is 60 seconds. |
| `read_timeout` | `float \| None` | No | 60.0 | The time in seconds till a timeout exception is thrown when attempting to read from a connection.The default is 60 seconds. |
| `session_ttl` | `int \| None` | No | 3600 | The time in seconds till a session expires. The default is 3600 seconds (1 hour). |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `region_name` | `str` | No | us-east-2 | AWS Region for the Bedrock Runtime endpoint |
## Sample Configuration
```yaml
{}
api_key: ${env.AWS_BEARER_TOKEN_BEDROCK:=}
region_name: ${env.AWS_DEFAULT_REGION:=us-east-2}
```

10
docs/docs/providers/inference/remote_cerebras.mdx
View file

@ -14,14 +14,14 @@ Cerebras inference provider for running models on Cerebras Cloud platform.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `<class 'str'>` | No | https://api.cerebras.ai | Base URL for the Cerebras API |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://api.cerebras.ai/v1 | Base URL for the Cerebras API |
## Sample Configuration
```yaml
base_url: https://api.cerebras.ai
base_url: https://api.cerebras.ai/v1
api_key: ${env.CEREBRAS_API_KEY:=}
```

10
docs/docs/providers/inference/remote_databricks.mdx
View file

@ -14,14 +14,14 @@ Databricks inference provider for running models on Databricks' unified analytic
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_token` | `pydantic.types.SecretStr \| None` | No | | The Databricks API token |
| `url` | `str \| None` | No | | The URL for the Databricks model serving endpoint |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_token` | `SecretStr \| None` | No | | The Databricks API token |
| `base_url` | `HttpUrl \| None` | No | | The URL for the Databricks model serving endpoint (should include /serving-endpoints path) |
## Sample Configuration
```yaml
url: ${env.DATABRICKS_HOST:=}
base_url: ${env.DATABRICKS_HOST:=}
api_token: ${env.DATABRICKS_TOKEN:=}
```

10
docs/docs/providers/inference/remote_fireworks.mdx
View file

@ -14,14 +14,14 @@ Fireworks AI inference provider for Llama models and other AI models on the Fire
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `url` | `<class 'str'>` | No | https://api.fireworks.ai/inference/v1 | The URL for the Fireworks server |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://api.fireworks.ai/inference/v1 | The URL for the Fireworks server |
## Sample Configuration
```yaml
url: https://api.fireworks.ai/inference/v1
base_url: https://api.fireworks.ai/inference/v1
api_key: ${env.FIREWORKS_API_KEY:=}
```

6
docs/docs/providers/inference/remote_gemini.mdx
View file

@ -14,9 +14,9 @@ Google Gemini inference provider for accessing Gemini models and Google's AI ser
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
## Sample Configuration

10
docs/docs/providers/inference/remote_groq.mdx
View file

@ -14,14 +14,14 @@ Groq inference provider for ultra-fast inference using Groq's LPU technology.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `url` | `<class 'str'>` | No | https://api.groq.com | The URL for the Groq AI server |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://api.groq.com/openai/v1 | The URL for the Groq AI server |
## Sample Configuration
```yaml
url: https://api.groq.com
base_url: https://api.groq.com/openai/v1
api_key: ${env.GROQ_API_KEY:=}
```

4
docs/docs/providers/inference/remote_hf_endpoint.mdx
View file

@ -14,8 +14,8 @@ HuggingFace Inference Endpoints provider for dedicated model serving.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `endpoint_name` | `<class 'str'>` | No | | The name of the Hugging Face Inference Endpoint in the format of '&#123;namespace&#125;/&#123;endpoint_name&#125;' (e.g. 'my-cool-org/meta-llama-3-1-8b-instruct-rce'). Namespace is optional and will default to the user account if not provided. |
| `api_token` | `pydantic.types.SecretStr \| None` | No | | Your Hugging Face user access token (will default to locally saved token if not provided) |
| `endpoint_name` | `str` | No | | The name of the Hugging Face Inference Endpoint in the format of '&#123;namespace&#125;/&#123;endpoint_name&#125;' (e.g. 'my-cool-org/meta-llama-3-1-8b-instruct-rce'). Namespace is optional and will default to the user account if not provided. |
| `api_token` | `SecretStr \| None` | No | | Your Hugging Face user access token (will default to locally saved token if not provided) |
## Sample Configuration

4
docs/docs/providers/inference/remote_hf_serverless.mdx
View file

@ -14,8 +14,8 @@ HuggingFace Inference API serverless provider for on-demand model inference.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `huggingface_repo` | `<class 'str'>` | No | | The model ID of the model on the Hugging Face Hub (e.g. 'meta-llama/Meta-Llama-3.1-70B-Instruct') |
| `api_token` | `pydantic.types.SecretStr \| None` | No | | Your Hugging Face user access token (will default to locally saved token if not provided) |
| `huggingface_repo` | `str` | No | | The model ID of the model on the Hugging Face Hub (e.g. 'meta-llama/Meta-Llama-3.1-70B-Instruct') |
| `api_token` | `SecretStr \| None` | No | | Your Hugging Face user access token (will default to locally saved token if not provided) |
## Sample Configuration

10
docs/docs/providers/inference/remote_llama-openai-compat.mdx
View file

@ -14,14 +14,14 @@ Llama OpenAI-compatible provider for using Llama models with OpenAI API format.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `openai_compat_api_base` | `<class 'str'>` | No | https://api.llama.com/compat/v1/ | The URL for the Llama API server |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://api.llama.com/compat/v1/ | The URL for the Llama API server |
## Sample Configuration
```yaml
openai_compat_api_base: https://api.llama.com/compat/v1/
base_url: https://api.llama.com/compat/v1/
api_key: ${env.LLAMA_API_KEY}
```

15
docs/docs/providers/inference/remote_nvidia.mdx
View file

@ -14,17 +14,16 @@ NVIDIA inference provider for accessing NVIDIA NIM models and AI services.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `url` | `<class 'str'>` | No | https://integrate.api.nvidia.com | A base url for accessing the NVIDIA NIM |
| `timeout` | `<class 'int'>` | No | 60 | Timeout for the HTTP requests |
| `append_api_version` | `<class 'bool'>` | No | True | When set to false, the API version will not be appended to the base_url. By default, it is true. |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://integrate.api.nvidia.com/v1 | A base url for accessing the NVIDIA NIM |
| `timeout` | `int` | No | 60 | Timeout for the HTTP requests |
| `rerank_model_to_url` | `dict[str, str]` | No | `{'nv-rerank-qa-mistral-4b:1': 'https://ai.api.nvidia.com/v1/retrieval/nvidia/reranking', 'nvidia/nv-rerankqa-mistral-4b-v3': 'https://ai.api.nvidia.com/v1/retrieval/nvidia/nv-rerankqa-mistral-4b-v3/reranking', 'nvidia/llama-3.2-nv-rerankqa-1b-v2': 'https://ai.api.nvidia.com/v1/retrieval/nvidia/llama-3_2-nv-rerankqa-1b-v2/reranking'}` | Mapping of rerank model identifiers to their API endpoints. |
## Sample Configuration
```yaml
url: ${env.NVIDIA_BASE_URL:=https://integrate.api.nvidia.com}
base_url: ${env.NVIDIA_BASE_URL:=https://integrate.api.nvidia.com/v1}
api_key: ${env.NVIDIA_API_KEY:=}
append_api_version: ${env.NVIDIA_APPEND_API_VERSION:=True}
```

41
docs/docs/providers/inference/remote_oci.mdx Normal file
View file

@ -0,0 +1,41 @@
---
description: |
Oracle Cloud Infrastructure (OCI) Generative AI inference provider for accessing OCI's Generative AI Platform-as-a-Service models.
Provider documentation
https://docs.oracle.com/en-us/iaas/Content/generative-ai/home.htm
sidebar_label: Remote - Oci
title: remote::oci
---
# remote::oci
## Description
Oracle Cloud Infrastructure (OCI) Generative AI inference provider for accessing OCI's Generative AI Platform-as-a-Service models.
Provider documentation
https://docs.oracle.com/en-us/iaas/Content/generative-ai/home.htm
## Configuration
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `oci_auth_type` | `str` | No | instance_principal | OCI authentication type (must be one of: instance_principal, config_file) |
| `oci_region` | `str` | No | us-ashburn-1 | OCI region (e.g., us-ashburn-1) |
| `oci_compartment_id` | `str` | No | | OCI compartment ID for the Generative AI service |
| `oci_config_file_path` | `str` | No | ~/.oci/config | OCI config file path (required if oci_auth_type is config_file) |
| `oci_config_profile` | `str` | No | DEFAULT | OCI config profile (required if oci_auth_type is config_file) |
## Sample Configuration
```yaml
oci_auth_type: ${env.OCI_AUTH_TYPE:=instance_principal}
oci_config_file_path: ${env.OCI_CONFIG_FILE_PATH:=~/.oci/config}
oci_config_profile: ${env.OCI_CLI_PROFILE:=DEFAULT}
oci_region: ${env.OCI_REGION:=us-ashburn-1}
oci_compartment_id: ${env.OCI_COMPARTMENT_OCID:=}
```

8
docs/docs/providers/inference/remote_ollama.mdx
View file

@ -14,12 +14,12 @@ Ollama inference provider for running local models through the Ollama runtime.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `url` | `<class 'str'>` | No | http://localhost:11434 | |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `base_url` | `HttpUrl \| None` | No | http://localhost:11434/v1 | |
## Sample Configuration
```yaml
url: ${env.OLLAMA_URL:=http://localhost:11434}
base_url: ${env.OLLAMA_URL:=http://localhost:11434/v1}
```

8
docs/docs/providers/inference/remote_openai.mdx
View file

@ -14,10 +14,10 @@ OpenAI inference provider for accessing GPT models and other OpenAI services.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `<class 'str'>` | No | https://api.openai.com/v1 | Base URL for OpenAI API |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://api.openai.com/v1 | Base URL for OpenAI API |
## Sample Configuration

10
docs/docs/providers/inference/remote_passthrough.mdx
View file

@ -14,14 +14,14 @@ Passthrough inference provider for connecting to any external inference service
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | API Key for the passthrouth endpoint |
| `url` | `<class 'str'>` | No | | The URL for the passthrough endpoint |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | | The URL for the passthrough endpoint |
## Sample Configuration
```yaml
url: ${env.PASSTHROUGH_URL}
base_url: ${env.PASSTHROUGH_URL}
api_key: ${env.PASSTHROUGH_API_KEY}
```

10
docs/docs/providers/inference/remote_runpod.mdx
View file

@ -14,14 +14,14 @@ RunPod inference provider for running models on RunPod's cloud GPU platform.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_token` | `pydantic.types.SecretStr \| None` | No | | The API token |
| `url` | `str \| None` | No | | The URL for the Runpod model serving endpoint |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_token` | `SecretStr \| None` | No | | The API token |
| `base_url` | `HttpUrl \| None` | No | | The URL for the Runpod model serving endpoint |
## Sample Configuration
```yaml
url: ${env.RUNPOD_URL:=}
base_url: ${env.RUNPOD_URL:=}
api_token: ${env.RUNPOD_API_TOKEN}
```

10
docs/docs/providers/inference/remote_sambanova.mdx
View file

@ -14,14 +14,14 @@ SambaNova inference provider for running models on SambaNova's dataflow architec
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `url` | `<class 'str'>` | No | https://api.sambanova.ai/v1 | The URL for the SambaNova AI server |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://api.sambanova.ai/v1 | The URL for the SambaNova AI server |
## Sample Configuration
```yaml
url: https://api.sambanova.ai/v1
base_url: https://api.sambanova.ai/v1
api_key: ${env.SAMBANOVA_API_KEY:=}
```

8
docs/docs/providers/inference/remote_tgi.mdx
View file

@ -14,12 +14,12 @@ Text Generation Inference (TGI) provider for HuggingFace model serving.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `url` | `<class 'str'>` | No | | The URL for the TGI serving endpoint |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `base_url` | `HttpUrl \| None` | No | | The URL for the TGI serving endpoint (should include /v1 path) |
## Sample Configuration
```yaml
url: ${env.TGI_URL:=}
base_url: ${env.TGI_URL:=}
```

10
docs/docs/providers/inference/remote_together.mdx
View file

@ -14,14 +14,14 @@ Together AI inference provider for open-source models and collaborative AI devel
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `url` | `<class 'str'>` | No | https://api.together.xyz/v1 | The URL for the Together AI server |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://api.together.xyz/v1 | The URL for the Together AI server |
## Sample Configuration
```yaml
url: https://api.together.xyz/v1
base_url: https://api.together.xyz/v1
api_key: ${env.TOGETHER_API_KEY:=}
```

8
docs/docs/providers/inference/remote_vertexai.mdx
View file

@ -53,10 +53,10 @@ Available Models:
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `project` | `<class 'str'>` | No | | Google Cloud project ID for Vertex AI |
| `location` | `<class 'str'>` | No | us-central1 | Google Cloud location for Vertex AI |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `project` | `str` | No | | Google Cloud project ID for Vertex AI |
| `location` | `str` | No | us-central1 | Google Cloud location for Vertex AI |
## Sample Configuration

12
docs/docs/providers/inference/remote_vllm.mdx
View file

@ -14,17 +14,17 @@ Remote vLLM inference provider for connecting to vLLM servers.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_token` | `pydantic.types.SecretStr \| None` | No | | The API token |
| `url` | `str \| None` | No | | The URL for the vLLM model serving endpoint |
| `max_tokens` | `<class 'int'>` | No | 4096 | Maximum number of tokens to generate. |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_token` | `SecretStr \| None` | No | | The API token |
| `base_url` | `HttpUrl \| None` | No | | The URL for the vLLM model serving endpoint |
| `max_tokens` | `int` | No | 4096 | Maximum number of tokens to generate. |
| `tls_verify` | `bool \| str` | No | True | Whether to verify TLS certificates. Can be a boolean or a path to a CA certificate file. |
## Sample Configuration
```yaml
url: ${env.VLLM_URL:=}
base_url: ${env.VLLM_URL:=}
max_tokens: ${env.VLLM_MAX_TOKENS:=4096}
api_token: ${env.VLLM_API_TOKEN:=fake}
tls_verify: ${env.VLLM_TLS_VERIFY:=true}

12
docs/docs/providers/inference/remote_watsonx.mdx
View file

@ -14,17 +14,17 @@ IBM WatsonX inference provider for accessing AI models on IBM's WatsonX platform
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | Authentication credential for the provider |
| `url` | `<class 'str'>` | No | https://us-south.ml.cloud.ibm.com | A base url for accessing the watsonx.ai |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `api_key` | `SecretStr \| None` | No | | Authentication credential for the provider |
| `base_url` | `HttpUrl \| None` | No | https://us-south.ml.cloud.ibm.com | A base url for accessing the watsonx.ai |
| `project_id` | `str \| None` | No | | The watsonx.ai project ID |
| `timeout` | `<class 'int'>` | No | 60 | Timeout for the HTTP requests |
| `timeout` | `int` | No | 60 | Timeout for the HTTP requests |
## Sample Configuration
```yaml
url: ${env.WATSONX_BASE_URL:=https://us-south.ml.cloud.ibm.com}
base_url: ${env.WATSONX_BASE_URL:=https://us-south.ml.cloud.ibm.com}
api_key: ${env.WATSONX_API_KEY:=}
project_id: ${env.WATSONX_PROJECT_ID:=}
```

19
docs/docs/providers/openai.mdx
View file

@ -1,9 +1,14 @@
---
title: OpenAI Compatibility
description: OpenAI API Compatibility
sidebar_label: OpenAI Compatibility
sidebar_position: 1
title: OpenAI Implementation Guide
description: Code examples and implementation details for OpenAI API compatibility
sidebar_label: OpenAI Implementation
sidebar_position: 2
---
# OpenAI Implementation Guide
This guide provides detailed code examples and implementation details for using OpenAI-compatible APIs with Llama Stack. For a comprehensive overview of OpenAI compatibility features, see our [OpenAI API Compatibility Guide](../api-openai/index.mdx).
## OpenAI API Compatibility
### Server path
@ -195,3 +200,9 @@ Lines of code unfurl
Logic whispers in the dark
Art in hidden form
```
## Additional Resources
- **[OpenAI API Compatibility Guide](../api-openai/index.mdx)** - Comprehensive overview of OpenAI compatibility features
- **[OpenAI Responses API Limitations](./openai_responses_limitations.mdx)** - Detailed limitations and known issues
- **[Provider Documentation](../index.mdx)** - Complete provider ecosystem overview

6
docs/docs/providers/openai_responses_limitations.mdx
View file

@ -48,11 +48,9 @@ Both OpenAI and Llama Stack support a web-search built-in tool. The [OpenAI doc
> The type of the web search tool. One of `web_search` or `web_search_2025_08_26`.
In contrast, the [Llama Stack documentation](https://llamastack.github.io/docs/api/create-a-new-open-ai-response) says that the allowed values for `type` for web search are `MOD1`, `MOD2` and `MOD3`.
Is that correct? If so, what are the meanings of each of them? It might make sense for the allowed values for OpenAI map to some values for Llama Stack so that code written to the OpenAI specification
also work with Llama Stack.
Llama Stack now supports both `web_search` and `web_search_2025_08_26` types, matching OpenAI's API. For backward compatibility, Llama Stack also supports `web_search_preview` and `web_search_preview_2025_03_11` types.
The OpenAI web search tool also has fields for `filters` and `user_location` which are not documented as options for Llama Stack. If feasible, it would be good to support these too.
The OpenAI web search tool also has fields for `filters` and `user_location` which are not yet implemented in Llama Stack. If feasible, it would be good to support these too.
---

34
docs/docs/providers/post_training/inline_huggingface-gpu.mdx
View file

@ -14,23 +14,23 @@ HuggingFace-based post-training provider for fine-tuning models using the Huggin
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `device` | `<class 'str'>` | No | cuda | |
| `distributed_backend` | `Literal['fsdp', 'deepspeed'` | No | | |
| `checkpoint_format` | `Literal['full_state', 'huggingface'` | No | huggingface | |
| `chat_template` | `<class 'str'>` | No | `&lt;|user|&gt;`<br/>`{input}`<br/>`&lt;|assistant|&gt;`<br/>`{output}` | |
| `model_specific_config` | `<class 'dict'>` | No | `{'trust_remote_code': True, 'attn_implementation': 'sdpa'}` | |
| `max_seq_length` | `<class 'int'>` | No | 2048 | |
| `gradient_checkpointing` | `<class 'bool'>` | No | False | |
| `save_total_limit` | `<class 'int'>` | No | 3 | |
| `logging_steps` | `<class 'int'>` | No | 10 | |
| `warmup_ratio` | `<class 'float'>` | No | 0.1 | |
| `weight_decay` | `<class 'float'>` | No | 0.01 | |
| `dataloader_num_workers` | `<class 'int'>` | No | 4 | |
| `dataloader_pin_memory` | `<class 'bool'>` | No | True | |
| `dpo_beta` | `<class 'float'>` | No | 0.1 | |
| `use_reference_model` | `<class 'bool'>` | No | True | |
| `dpo_loss_type` | `Literal['sigmoid', 'hinge', 'ipo', 'kto_pair'` | No | sigmoid | |
| `dpo_output_dir` | `<class 'str'>` | No | | |
| `device` | `str` | No | cuda | |
| `distributed_backend` | `Literal[fsdp, deepspeed] \| None` | No | | |
| `checkpoint_format` | `Literal[full_state, huggingface] \| None` | No | huggingface | |
| `chat_template` | `str` | No | `&lt;|user|&gt;`<br/>`{input}`<br/>`&lt;|assistant|&gt;`<br/>`{output}` | |
| `model_specific_config` | `dict` | No | `{'trust_remote_code': True, 'attn_implementation': 'sdpa'}` | |
| `max_seq_length` | `int` | No | 2048 | |
| `gradient_checkpointing` | `bool` | No | False | |
| `save_total_limit` | `int` | No | 3 | |
| `logging_steps` | `int` | No | 10 | |
| `warmup_ratio` | `float` | No | 0.1 | |
| `weight_decay` | `float` | No | 0.01 | |
| `dataloader_num_workers` | `int` | No | 4 | |
| `dataloader_pin_memory` | `bool` | No | True | |
| `dpo_beta` | `float` | No | 0.1 | |
| `use_reference_model` | `bool` | No | True | |
| `dpo_loss_type` | `Literal[sigmoid, hinge, ipo, kto_pair]` | No | sigmoid | |
| `dpo_output_dir` | `str` | No | | |
## Sample Configuration

2
docs/docs/providers/post_training/inline_torchtune-cpu.mdx
View file

@ -15,7 +15,7 @@ TorchTune-based post-training provider for fine-tuning and optimizing models usi
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `torch_seed` | `int \| None` | No | | |
| `checkpoint_format` | `Literal['meta', 'huggingface'` | No | meta | |
| `checkpoint_format` | `Literal[meta, huggingface] \| None` | No | meta | |
## Sample Configuration

2
docs/docs/providers/post_training/inline_torchtune-gpu.mdx
View file

@ -15,7 +15,7 @@ TorchTune-based post-training provider for fine-tuning and optimizing models usi
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `torch_seed` | `int \| None` | No | | |
| `checkpoint_format` | `Literal['meta', 'huggingface'` | No | meta | |
| `checkpoint_format` | `Literal[meta, huggingface] \| None` | No | meta | |
## Sample Configuration

6
docs/docs/providers/post_training/remote_nvidia.mdx
View file

@ -18,9 +18,9 @@ NVIDIA's post-training provider for fine-tuning models on NVIDIA's platform.
| `dataset_namespace` | `str \| None` | No | default | The NVIDIA dataset namespace. |
| `project_id` | `str \| None` | No | test-example-model@v1 | The NVIDIA project ID. |
| `customizer_url` | `str \| None` | No | | Base URL for the NeMo Customizer API |
| `timeout` | `<class 'int'>` | No | 300 | Timeout for the NVIDIA Post Training API |
| `max_retries` | `<class 'int'>` | No | 3 | Maximum number of retries for the NVIDIA Post Training API |
| `output_model_dir` | `<class 'str'>` | No | test-example-model@v1 | Directory to save the output model |
| `timeout` | `int` | No | 300 | Timeout for the NVIDIA Post Training API |
| `max_retries` | `int` | No | 3 | Maximum number of retries for the NVIDIA Post Training API |
| `output_model_dir` | `str` | No | test-example-model@v1 | Directory to save the output model |
## Sample Configuration

7
docs/docs/providers/safety/index.mdx
View file

@ -1,7 +1,8 @@
---
description: "Safety
description: |
Safety
OpenAI-compatible Moderations API."
OpenAI-compatible Moderations API.
sidebar_label: Safety
title: Safety
---
@ -12,6 +13,6 @@ title: Safety
Safety
OpenAI-compatible Moderations API.
OpenAI-compatible Moderations API.
This section contains documentation for all available providers for the **safety** API.

2
docs/docs/providers/safety/inline_llama-guard.mdx
View file

@ -14,7 +14,7 @@ Llama Guard safety provider for content moderation and safety filtering using Me
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `excluded_categories` | `list[str` | No | [] | |
| `excluded_categories` | `list[str]` | No | [] | |
## Sample Configuration

2
docs/docs/providers/safety/inline_prompt-guard.mdx
View file

@ -14,7 +14,7 @@ Prompt Guard safety provider for detecting and filtering unsafe prompts and cont
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `guard_type` | `<class 'str'>` | No | injection | |
| `guard_type` | `str` | No | injection | |
## Sample Configuration

4
docs/docs/providers/safety/remote_bedrock.mdx
View file

@ -14,8 +14,8 @@ AWS Bedrock safety provider for content moderation using AWS's safety services.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `allowed_models` | `list[str \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `<class 'bool'>` | No | False | Whether to refresh models periodically from the provider |
| `allowed_models` | `list[str] \| None` | No | | List of models that should be registered with the model registry. If None, all models are allowed. |
| `refresh_models` | `bool` | No | False | Whether to refresh models periodically from the provider |
| `aws_access_key_id` | `str \| None` | No | | The AWS access key to use. Default use environment variable: AWS_ACCESS_KEY_ID |
| `aws_secret_access_key` | `str \| None` | No | | The AWS secret access key to use. Default use environment variable: AWS_SECRET_ACCESS_KEY |
| `aws_session_token` | `str \| None` | No | | The AWS session token to use. Default use environment variable: AWS_SESSION_TOKEN |

2
docs/docs/providers/safety/remote_nvidia.mdx
View file

@ -14,7 +14,7 @@ NVIDIA's safety provider for content moderation and safety filtering.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `guardrails_service_url` | `<class 'str'>` | No | http://0.0.0.0:7331 | The url for accessing the Guardrails service |
| `guardrails_service_url` | `str` | No | http://0.0.0.0:7331 | The url for accessing the Guardrails service |
| `config_id` | `str \| None` | No | self-check | Guardrails configuration ID to use from the Guardrails configuration store |
## Sample Configuration

4
docs/docs/providers/safety/remote_sambanova.mdx
View file

@ -14,8 +14,8 @@ SambaNova's safety provider for content moderation and safety filtering.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `url` | `<class 'str'>` | No | https://api.sambanova.ai/v1 | The URL for the SambaNova AI server |
| `api_key` | `pydantic.types.SecretStr \| None` | No | | The SambaNova cloud API Key |
| `url` | `str` | No | https://api.sambanova.ai/v1 | The URL for the SambaNova AI server |
| `api_key` | `SecretStr \| None` | No | | The SambaNova cloud API Key |
## Sample Configuration

10
docs/docs/providers/telemetry/index.mdx
View file

@ -1,10 +0,0 @@
---
sidebar_label: Telemetry
title: Telemetry
---
# Telemetry
## Overview
This section contains documentation for all available providers for the **telemetry** API.

27
docs/docs/providers/telemetry/inline_meta-reference.mdx
View file

@ -1,27 +0,0 @@
---
description: "Meta's reference implementation of telemetry and observability using OpenTelemetry."
sidebar_label: Meta-Reference
title: inline::meta-reference
---
# inline::meta-reference
## Description
Meta's reference implementation of telemetry and observability using OpenTelemetry.
## Configuration
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `otel_exporter_otlp_endpoint` | `str \| None` | No | | The OpenTelemetry collector endpoint URL (base URL for traces, metrics, and logs). If not set, the SDK will use OTEL_EXPORTER_OTLP_ENDPOINT environment variable. |
| `service_name` | `<class 'str'>` | No | ​ | The service name to use for telemetry |
| `sinks` | `list[inline.telemetry.meta_reference.config.TelemetrySink` | No | [] | List of telemetry sinks to enable (possible values: otel_trace, otel_metric, console) |
## Sample Configuration
```yaml
service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
sinks: ${env.TELEMETRY_SINKS:=}
otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
```

2
docs/docs/providers/tool_runtime/remote_bing-search.mdx
View file

@ -15,7 +15,7 @@ Bing Search tool for web search capabilities using Microsoft's search engine.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `api_key` | `str \| None` | No | | |
| `top_k` | `<class 'int'>` | No | 3 | |
| `top_k` | `int` | No | 3 | |
## Sample Configuration

2
docs/docs/providers/tool_runtime/remote_brave-search.mdx
View file

@ -15,7 +15,7 @@ Brave Search tool for web search capabilities with privacy-focused results.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `api_key` | `str \| None` | No | | The Brave Search API Key |
| `max_results` | `<class 'int'>` | No | 3 | The maximum number of results to return |
| `max_results` | `int` | No | 3 | The maximum number of results to return |
## Sample Configuration

2
docs/docs/providers/tool_runtime/remote_tavily-search.mdx
View file

@ -15,7 +15,7 @@ Tavily Search tool for AI-optimized web search with structured results.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `api_key` | `str \| None` | No | | The Tavily Search API Key |
| `max_results` | `<class 'int'>` | No | 3 | The maximum number of results to return |
| `max_results` | `int` | No | 3 | The maximum number of results to return |
## Sample Configuration

4
docs/docs/providers/vector_io/inline_chromadb.mdx
View file

@ -78,8 +78,8 @@ See [Chroma's documentation](https://docs.trychroma.com/docs/overview/introducti
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `db_path` | `<class 'str'>` | No | | |
| `persistence` | `<class 'llama_stack.core.storage.datatypes.KVStoreReference'>` | No | | Config for KV store backend |
| `db_path` | `str` | No | | |
| `persistence` | `KVStoreReference` | No | | Config for KV store backend |
## Sample Configuration

2
docs/docs/providers/vector_io/inline_faiss.mdx
View file

@ -95,7 +95,7 @@ more details about Faiss in general.
| Field | Type | Required | Default | Description |
|-------|------|----------|---------|-------------|
| `persistence` | `<class 'llama_stack.core.storage.datatypes.KVStoreReference'>` | No | | |
| `persistence` | `KVStoreReference` | No | | |
## Sample Configuration

Some files were not shown because too many files have changed in this diff Show more