feat: split API and provider specs into separate llama-stack-api pkg

Extract API definitions, models, and provider specifications into a
standalone llama-stack-api package that can be published to PyPI
independently of the main llama-stack server.

Motivation

External providers currently import from llama-stack, which overrides
the installed version and causes dependency conflicts. This separation
allows external providers to:

- Install only the type definitions they need without server dependencies
- Avoid version conflicts with the installed llama-stack package
- Be versioned and released independently

This enables us to re-enable external provider module tests that were
previously blocked by these import conflicts.

Changes

- Created llama-stack-api package with minimal dependencies (pydantic, jsonschema)
- Moved APIs, providers datatypes, strong_typing, and schema_utils
- Updated all imports from llama_stack.* to llama_stack_api.*
- Preserved git history using git mv for moved files
- Configured local editable install for development workflow
- Updated linting and type-checking configuration for both packages
- Rebased on top of upstream src/ layout changes

Testing

Package builds successfully and can be imported independently.
All pre-commit hooks pass with expected exclusions maintained.

Next Steps

- Publish llama-stack-api to PyPI
- Update external provider dependencies
- Re-enable external provider module tests

Signed-off-by: Charlie Doern <cdoern@redhat.com>

.github/workflows/python-build-test.yml

@@ -30,13 +30,16 @@ jobs:
         activate-environment: true
         version: 0.7.6
 
-    - name: Build Llama Stack package
-      run: |
-        uv build
+    - name: Build Llama Stack Spec package
+      working-directory: src/llama-stack-api
+      run: uv build
 
-    - name: Install Llama Stack package
+    - name: Build Llama Stack package
+      run: uv build
+
+    - name: Install Llama Stack package (with spec from local build)
       run: |
-        uv pip install dist/*.whl
+        uv pip install --find-links src/llama-stack-api/dist dist/*.whl
 
     - name: Verify Llama Stack package
       run: |

.pre-commit-config.yaml

@@ -42,7 +42,7 @@ repos:
     hooks:
     -   id: ruff
         args: [ --fix ]
-        exclude: ^src/llama_stack/strong_typing/.*$
+        exclude: ^(src/llama_stack/strong_typing/.*|src/llama-stack-api/llama_stack_api/strong_typing/.*)$
     -   id: ruff-format
 
 -   repo: https://github.com/adamchainz/blacken-docs

docs/docs/concepts/apis/external.mdx

@@ -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.providers.datatypes 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.schema_utils import webmethod
 
 
 class WeatherAPI(Protocol):
@@ -151,12 +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.providers.datatypes import (
     Api,
     ProviderSpec,
     RemoteProviderSpec,
 )
-from llama_stack.schema_utils import webmethod
+from llama_stack_api.schema_utils import webmethod
 
 
 def available_providers() -> list[ProviderSpec]:

docs/docs/distributions/building_distro.mdx

@@ -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:

docs/docs/providers/external/external-providers-guide.mdx

@@ -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,

docs/docs/providers/vector_io/inline_sqlite-vec.mdx

@@ -153,7 +153,7 @@ description: |
   Example using RAGQueryConfig with different search modes:
 
   ```python
-  from llama_stack.apis.tools import RAGQueryConfig, RRFRanker, WeightedRanker
+  from llama_stack_api.apis.tools import RAGQueryConfig, RRFRanker, WeightedRanker
 
   # Vector search
   config = RAGQueryConfig(mode="vector", max_chunks=5)
@@ -358,7 +358,7 @@ Two ranker types are supported:
 Example using RAGQueryConfig with different search modes:
 
 ```python
-from llama_stack.apis.tools import RAGQueryConfig, RRFRanker, WeightedRanker
+from llama_stack_api.apis.tools import RAGQueryConfig, RRFRanker, WeightedRanker
 
 # Vector search
 config = RAGQueryConfig(mode="vector", max_chunks=5)

docs/openapi_generator/generate.py

@@ -16,7 +16,7 @@ import sys
 import fire
 import ruamel.yaml as yaml
 
-from llama_stack.apis.version import LLAMA_STACK_API_V1 # noqa: E402
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1 # noqa: E402
 from llama_stack.core.stack import LlamaStack  # noqa: E402
 
 from .pyopenapi.options import Options  # noqa: E402

docs/openapi_generator/pyopenapi/generator.py

@@ -16,10 +16,10 @@ from typing import Annotated, Any, Dict, get_args, get_origin, Set, Union
 
 from fastapi import UploadFile
 
-from llama_stack.apis.datatypes import Error
-from llama_stack.strong_typing.core import JsonType
-from llama_stack.strong_typing.docstring import Docstring, parse_type
-from llama_stack.strong_typing.inspection import (
+from llama_stack_api.apis.datatypes import Error
+from llama_stack_api.strong_typing.core import JsonType
+from llama_stack_api.strong_typing.docstring import Docstring, parse_type
+from llama_stack_api.strong_typing.inspection import (
     is_generic_list,
     is_type_optional,
     is_type_union,
@@ -28,15 +28,15 @@ from llama_stack.strong_typing.inspection import (
     unwrap_optional_type,
     unwrap_union_types,
 )
-from llama_stack.strong_typing.name import python_type_to_name
-from llama_stack.strong_typing.schema import (
+from llama_stack_api.strong_typing.name import python_type_to_name
+from llama_stack_api.strong_typing.schema import (
     get_schema_identifier,
     JsonSchemaGenerator,
     register_schema,
     Schema,
     SchemaOptions,
 )
-from llama_stack.strong_typing.serialization import json_dump_string, object_to_json
+from llama_stack_api.strong_typing.serialization import json_dump_string, object_to_json
 from pydantic import BaseModel
 
 from .operations import (

docs/openapi_generator/pyopenapi/operations.py

@@ -11,11 +11,11 @@ import typing
 from dataclasses import dataclass
 from typing import Any, Callable, Dict, Iterable, Iterator, List, Optional, Tuple, Union
 
-from llama_stack.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1BETA, LLAMA_STACK_API_V1ALPHA
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1BETA, LLAMA_STACK_API_V1ALPHA
 
 from termcolor import colored
 
-from llama_stack.strong_typing.inspection import get_signature
+from llama_stack_api.strong_typing.inspection import get_signature
 
 from typing import get_origin, get_args
 
@@ -23,7 +23,7 @@ from fastapi import UploadFile
 from fastapi.params import File, Form
 from typing import Annotated
 
-from llama_stack.schema_utils import ExtraBodyField
+from llama_stack_api.schema_utils import ExtraBodyField
 
 
 def split_prefix(

docs/openapi_generator/pyopenapi/specification.py

@@ -9,7 +9,7 @@ import enum
 from dataclasses import dataclass
 from typing import Any, ClassVar, Dict, List, Optional, Union
 
-from llama_stack.strong_typing.schema import JsonType, Schema, StrictJsonType
+from llama_stack_api.strong_typing.schema import JsonType, Schema, StrictJsonType
 
 URL = str
 

docs/openapi_generator/pyopenapi/utility.py

@@ -11,8 +11,8 @@ from pathlib import Path
 from typing import Any, List, Optional, TextIO, Union, get_type_hints, get_origin, get_args
 
 from pydantic import BaseModel
-from llama_stack.strong_typing.schema import object_to_json, StrictJsonType
-from llama_stack.strong_typing.inspection import is_unwrapped_body_param
+from llama_stack_api.strong_typing.schema import object_to_json, StrictJsonType
+from llama_stack_api.strong_typing.inspection import is_unwrapped_body_param
 from llama_stack.core.resolver import api_protocol_map
 
 from .generator import Generator

pyproject.toml

@@ -31,6 +31,7 @@ dependencies = [
     "httpx",
     "jinja2>=3.1.6",
     "jsonschema",
+    "llama-stack-api",  # API and provider specifications (local dev via tool.uv.sources)
     "openai>=2.5.0",
     "prompt-toolkit",
     "python-dotenv",
@@ -180,7 +181,7 @@ install-wheel-from-presigned = "llama_stack.cli.scripts.run:install_wheel_from_p
 
 [tool.setuptools.packages.find]
 where = ["src"]
-include = ["llama_stack", "llama_stack.*"]
+include = ["llama_stack", "llama_stack.*", "llama-stack-api", "llama-stack-api.*"]
 
 [[tool.uv.index]]
 name = "pytorch-cpu"
@@ -190,6 +191,7 @@ explicit = true
 [tool.uv.sources]
 torch = [{ index = "pytorch-cpu" }]
 torchvision = [{ index = "pytorch-cpu" }]
+llama-stack-api = [{ path = "src/llama-stack-api", editable = true }]
 
 [tool.ruff]
 line-length = 120
@@ -256,8 +258,8 @@ unfixable = [
 ] # Using import * is acceptable (or at least tolerated) in an __init__.py of a package API
 
 [tool.mypy]
-mypy_path = ["src"]
-packages = ["llama_stack"]
+mypy_path = ["src", "src/llama-stack-api"]
+packages = ["llama_stack", "llama_stack_api"]
 plugins = ['pydantic.mypy']
 disable_error_code = []
 warn_return_any = true
@@ -279,15 +281,18 @@ exclude = [
     "^src/llama_stack/core/store/registry\\.py$",
     "^src/llama_stack/core/utils/exec\\.py$",
     "^src/llama_stack/core/utils/prompt_for_config\\.py$",
+    # Moved to llama-stack-api but still excluded
     "^src/llama_stack/models/llama/llama3/interface\\.py$",
     "^src/llama_stack/models/llama/llama3/tokenizer\\.py$",
     "^src/llama_stack/models/llama/llama3/tool_utils\\.py$",
-    "^src/llama_stack/providers/inline/datasetio/localfs/",
-    "^src/llama_stack/providers/inline/eval/meta_reference/eval\\.py$",
-    "^src/llama_stack/providers/inline/inference/meta_reference/inference\\.py$",
     "^src/llama_stack/models/llama/llama3/generation\\.py$",
     "^src/llama_stack/models/llama/llama3/multimodal/model\\.py$",
     "^src/llama_stack/models/llama/llama4/",
+    "^src/llama-stack-api/llama_stack_api/core/telemetry/telemetry\\.py$",
+    "^src/llama_stack/providers/inline/agents/meta_reference/",
+    "^src/llama_stack/providers/inline/datasetio/localfs/",
+    "^src/llama_stack/providers/inline/eval/meta_reference/eval\\.py$",
+    "^src/llama_stack/providers/inline/inference/meta_reference/inference\\.py$",
     "^src/llama_stack/providers/inline/inference/sentence_transformers/sentence_transformers\\.py$",
     "^src/llama_stack/providers/inline/post_training/common/validator\\.py$",
     "^src/llama_stack/providers/inline/safety/code_scanner/",
@@ -337,7 +342,9 @@ exclude = [
     "^src/llama_stack/providers/utils/telemetry/dataset_mixin\\.py$",
     "^src/llama_stack/providers/utils/telemetry/trace_protocol\\.py$",
     "^src/llama_stack/providers/utils/telemetry/tracing\\.py$",
-    "^src/llama_stack/strong_typing/auxiliary\\.py$",
+    "^src/llama-stack-api/llama_stack_api/core/telemetry/trace_protocol\\.py$",
+    "^src/llama-stack-api/llama_stack_api/core/telemetry/tracing\\.py$",
+    "^src/llama-stack-api/llama_stack_api/strong_typing/auxiliary\\.py$",
     "^src/llama_stack/distributions/template\\.py$",
 ]
 

scripts/generate_prompt_format.py

@@ -14,8 +14,8 @@ import os
 from pathlib import Path
 
 import fire
+from llama_stack_api.apis.common.errors import ModelNotFoundError
 
-from llama_stack.apis.common.errors import ModelNotFoundError
 from llama_stack.models.llama.llama3.generation import Llama3
 from llama_stack.models.llama.llama4.generation import Llama4
 from llama_stack.models.llama.sku_list import resolve_model

scripts/provider_codegen.py

@@ -22,7 +22,7 @@ def get_api_docstring(api_name: str) -> str | None:
     """Extract docstring from the API protocol class."""
     try:
         # Import the API module dynamically
-        api_module = __import__(f"llama_stack.apis.{api_name}", fromlist=[api_name.title()])
+        api_module = __import__(f"llama_stack_api.apis.{api_name}", fromlist=[api_name.title()])
 
         # Get the main protocol class (usually capitalized API name)
         protocol_class_name = api_name.title()
@@ -83,8 +83,9 @@ def get_config_class_info(config_class_path: str) -> dict[str, Any]:
                 # this string replace is ridiculous
                 field_type = field_type.replace("typing.", "").replace("Optional[", "").replace("]", "")
                 field_type = field_type.replace("Annotated[", "").replace("FieldInfo(", "").replace(")", "")
-                field_type = field_type.replace("llama_stack.apis.inference.inference.", "")
+                field_type = field_type.replace("llama_stack_api.apis.inference.inference.", "")
                 field_type = field_type.replace("llama_stack.providers.", "")
+                field_type = field_type.replace("llama_stack_api.providers.", "")
 
                 default_value = field.default
                 if field.default_factory is not None:

src/llama-stack-api/README.md

@@ -0,0 +1,103 @@
+# llama-stack-api
+
+API and Provider specifications for Llama Stack - a lightweight package with protocol definitions and provider specs.
+
+## Overview
+
+`llama-stack-api` is a minimal dependency package that contains:
+
+- **API Protocol Definitions**: Type-safe protocol definitions for all Llama Stack APIs (inference, agents, safety, etc.)
+- **Provider Specifications**: Provider spec definitions for building custom providers
+- **Data Types**: Shared data types and models used across the Llama Stack ecosystem
+- **Type Utilities**: Strong typing utilities and schema validation
+
+## What This Package Does NOT Include
+
+- Server implementation (see `llama-stack` package)
+- Provider implementations (see `llama-stack` package)
+- CLI tools (see `llama-stack` package)
+- Runtime orchestration (see `llama-stack` package)
+
+## Use Cases
+
+This package is designed for:
+
+1. **Third-party Provider Developers**: Build custom providers without depending on the full Llama Stack server
+2. **Client Library Authors**: Use type definitions without server dependencies
+3. **Documentation Generation**: Generate API docs from protocol definitions
+4. **Type Checking**: Validate implementations against the official specs
+
+## Installation
+
+```bash
+pip install llama-stack-api
+```
+
+Or with uv:
+
+```bash
+uv pip install llama-stack-api
+```
+
+## Dependencies
+
+Minimal dependencies:
+- `pydantic>=2.11.9` - For data validation and serialization
+- `jsonschema` - For JSON schema utilities
+
+## Versioning
+
+This package follows semantic versioning independently from the main `llama-stack` package:
+
+- **Patch versions** (0.1.x): Documentation, internal improvements
+- **Minor versions** (0.x.0): New APIs, backward-compatible changes
+- **Major versions** (x.0.0): Breaking changes to existing APIs
+
+Current version: **0.1.0**
+
+## Usage Example
+
+```python
+from llama_stack_api.apis.inference import Inference, ChatCompletionRequest
+from llama_stack_api.providers.datatypes import ProviderSpec, InlineProviderSpec
+from llama_stack_api.apis.datatypes import Api
+
+
+# Use protocol definitions for type checking
+class MyInferenceProvider(Inference):
+    async def chat_completion(self, request: ChatCompletionRequest):
+        # Your implementation
+        pass
+
+
+# Define provider specifications
+my_provider_spec = InlineProviderSpec(
+    api=Api.inference,
+    provider_type="inline::my-provider",
+    pip_packages=["my-dependencies"],
+    module="my_package.providers.inference",
+    config_class="my_package.providers.inference.MyConfig",
+)
+```
+
+## Relationship to llama-stack
+
+The main `llama-stack` package depends on `llama-stack-api` and provides:
+- Full server implementation
+- Built-in provider implementations
+- CLI tools for running and managing stacks
+- Runtime provider resolution and orchestration
+
+## Contributing
+
+See the main [Llama Stack repository](https://github.com/llamastack/llama-stack) for contribution guidelines.
+
+## License
+
+MIT License - see LICENSE file for details.
+
+## Links
+
+- [Main Llama Stack Repository](https://github.com/llamastack/llama-stack)
+- [Documentation](https://llamastack.ai/)
+- [Client Library](https://pypi.org/project/llama-stack-client/)

src/llama-stack-api/llama_stack_api/__init__.py

@@ -0,0 +1,26 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+"""
+Llama Stack API Specifications
+
+This package contains the API definitions, data types, and protocol specifications
+for Llama Stack. It is designed to be a lightweight dependency for external providers
+and clients that need to interact with Llama Stack APIs without requiring the full
+server implementation.
+
+Key components:
+- apis: Protocol definitions for all Llama Stack APIs
+- providers: Provider interface specifications
+- strong_typing: Type system utilities
+- schema_utils: Schema validation and utilities
+"""
+
+__version__ = "0.1.0"
+
+from . import apis, providers, schema_utils, strong_typing  # noqa: F401
+
+__all__ = ["apis", "providers", "schema_utils", "strong_typing"]

src/llama-stack-api/llama_stack_api/apis/agents/agents.py

@@ -9,9 +9,9 @@ from typing import Annotated, Protocol, runtime_checkable
 
 from pydantic import BaseModel
 
-from llama_stack.apis.common.responses import Order
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import ExtraBodyField, json_schema_type, webmethod
+from llama_stack_api.apis.common.responses import Order
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import ExtraBodyField, json_schema_type, webmethod
 
 from .openai_responses import (
     ListOpenAIResponseInputItem,

src/llama-stack-api/llama_stack_api/apis/agents/openai_responses.py

@@ -10,8 +10,8 @@ from typing import Annotated, Any, Literal
 from pydantic import BaseModel, Field, model_validator
 from typing_extensions import TypedDict
 
-from llama_stack.apis.vector_io import SearchRankingOptions as FileSearchRankingOptions
-from llama_stack.schema_utils import json_schema_type, register_schema
+from llama_stack_api.apis.vector_io import SearchRankingOptions as FileSearchRankingOptions
+from llama_stack_api.schema_utils import json_schema_type, register_schema
 
 # NOTE(ashwin): this file is literally a copy of the OpenAI responses API schema. We should probably
 # take their YAML and generate this file automatically. Their YAML is available.

src/llama-stack-api/llama_stack_api/apis/batches/batches.py

@@ -8,8 +8,8 @@ from typing import Literal, Protocol, runtime_checkable
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 try:
     from openai.types import Batch as BatchObject

src/llama-stack-api/llama_stack_api/apis/benchmarks/benchmarks.py

@@ -7,9 +7,9 @@ from typing import Any, Literal, Protocol, runtime_checkable
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.resource import Resource, ResourceType
-from llama_stack.apis.version import LLAMA_STACK_API_V1ALPHA
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.resource import Resource, ResourceType
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1ALPHA
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 class CommonBenchmarkFields(BaseModel):

src/llama-stack-api/llama_stack_api/apis/common/content_types.py

@@ -4,13 +4,11 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
-from enum import Enum
 from typing import Annotated, Literal
 
 from pydantic import BaseModel, Field, model_validator
 
-from llama_stack.models.llama.datatypes import ToolCall
-from llama_stack.schema_utils import json_schema_type, register_schema
+from llama_stack_api.schema_utils import json_schema_type, register_schema
 
 
 @json_schema_type
@@ -101,43 +99,3 @@ class ImageDelta(BaseModel):
 
     type: Literal["image"] = "image"
     image: bytes
-
-
-class ToolCallParseStatus(Enum):
-    """Status of tool call parsing during streaming.
-    :cvar started: Tool call parsing has begun
-    :cvar in_progress: Tool call parsing is ongoing
-    :cvar failed: Tool call parsing failed
-    :cvar succeeded: Tool call parsing completed successfully
-    """
-
-    started = "started"
-    in_progress = "in_progress"
-    failed = "failed"
-    succeeded = "succeeded"
-
-
-@json_schema_type
-class ToolCallDelta(BaseModel):
-    """A tool call content delta for streaming responses.
-
-    :param type: Discriminator type of the delta. Always "tool_call"
-    :param tool_call: Either an in-progress tool call string or the final parsed tool call
-    :param parse_status: Current parsing status of the tool call
-    """
-
-    type: Literal["tool_call"] = "tool_call"
-
-    # you either send an in-progress tool call so the client can stream a long
-    # code generation or you send the final parsed tool call at the end of the
-    # stream
-    tool_call: str | ToolCall
-    parse_status: ToolCallParseStatus
-
-
-# streaming completions send a stream of ContentDeltas
-ContentDelta = Annotated[
-    TextDelta | ImageDelta | ToolCallDelta,
-    Field(discriminator="type"),
-]
-register_schema(ContentDelta, name="ContentDelta")

src/llama-stack-api/llama_stack_api/apis/common/job_types.py

@@ -7,7 +7,7 @@ from enum import Enum
 
 from pydantic import BaseModel
 
-from llama_stack.schema_utils import json_schema_type
+from llama_stack_api.schema_utils import json_schema_type
 
 
 class JobStatus(Enum):

src/llama-stack-api/llama_stack_api/apis/common/responses.py

@@ -9,7 +9,7 @@ from typing import Any
 
 from pydantic import BaseModel
 
-from llama_stack.schema_utils import json_schema_type
+from llama_stack_api.schema_utils import json_schema_type
 
 
 class Order(Enum):

src/llama-stack-api/llama_stack_api/apis/common/training_types.py

@@ -8,7 +8,7 @@ from datetime import datetime
 
 from pydantic import BaseModel
 
-from llama_stack.schema_utils import json_schema_type
+from llama_stack_api.schema_utils import json_schema_type
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/common/type_system.py

@@ -8,7 +8,7 @@ from typing import Annotated, Literal
 
 from pydantic import BaseModel, Field
 
-from llama_stack.schema_utils import json_schema_type, register_schema
+from llama_stack_api.schema_utils import json_schema_type, register_schema
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/conversations/conversations.py

@@ -9,7 +9,7 @@ from typing import Annotated, Literal, Protocol, runtime_checkable
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.agents.openai_responses import (
+from llama_stack_api.apis.agents.openai_responses import (
     OpenAIResponseInputFunctionToolCallOutput,
     OpenAIResponseMCPApprovalRequest,
     OpenAIResponseMCPApprovalResponse,
@@ -20,9 +20,9 @@ from llama_stack.apis.agents.openai_responses import (
     OpenAIResponseOutputMessageMCPListTools,
     OpenAIResponseOutputMessageWebSearchToolCall,
 )
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, register_schema, webmethod
 
 Metadata = dict[str, str]
 

src/llama-stack-api/llama_stack_api/apis/datasetio/datasetio.py

@@ -6,10 +6,10 @@
 
 from typing import Any, Protocol, runtime_checkable
 
-from llama_stack.apis.common.responses import PaginatedResponse
-from llama_stack.apis.datasets import Dataset
-from llama_stack.apis.version import LLAMA_STACK_API_V1BETA
-from llama_stack.schema_utils import webmethod
+from llama_stack_api.apis.common.responses import PaginatedResponse
+from llama_stack_api.apis.datasets import Dataset
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1BETA
+from llama_stack_api.schema_utils import webmethod
 
 
 class DatasetStore(Protocol):

src/llama-stack-api/llama_stack_api/apis/datasets/datasets.py

@@ -9,9 +9,9 @@ from typing import Annotated, Any, Literal, Protocol
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.resource import Resource, ResourceType
-from llama_stack.apis.version import LLAMA_STACK_API_V1BETA
-from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
+from llama_stack_api.apis.resource import Resource, ResourceType
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1BETA
+from llama_stack_api.schema_utils import json_schema_type, register_schema, webmethod
 
 
 class DatasetPurpose(StrEnum):

src/llama-stack-api/llama_stack_api/apis/datatypes.py

@@ -8,7 +8,7 @@ from enum import Enum, EnumMeta
 
 from pydantic import BaseModel, Field
 
-from llama_stack.schema_utils import json_schema_type
+from llama_stack_api.schema_utils import json_schema_type
 
 
 class DynamicApiMeta(EnumMeta):

src/llama-stack-api/llama_stack_api/apis/eval/eval.py

@@ -8,12 +8,12 @@ from typing import Any, Literal, Protocol
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.common.job_types import Job
-from llama_stack.apis.inference import SamplingParams, SystemMessage
-from llama_stack.apis.scoring import ScoringResult
-from llama_stack.apis.scoring_functions import ScoringFnParams
-from llama_stack.apis.version import LLAMA_STACK_API_V1ALPHA
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.common.job_types import Job
+from llama_stack_api.apis.inference import SamplingParams, SystemMessage
+from llama_stack_api.apis.scoring import ScoringResult
+from llama_stack_api.apis.scoring_functions import ScoringFnParams
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1ALPHA
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/files/files.py

@@ -10,10 +10,10 @@ from typing import Annotated, ClassVar, Literal, Protocol, runtime_checkable
 from fastapi import File, Form, Response, UploadFile
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.common.responses import Order
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.common.responses import Order
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 # OpenAI Files API Models

src/llama-stack-api/llama_stack_api/apis/inference/inference.py

@@ -18,14 +18,14 @@ from fastapi import Body
 from pydantic import BaseModel, Field
 from typing_extensions import TypedDict
 
-from llama_stack.apis.common.content_types import InterleavedContent
-from llama_stack.apis.common.responses import (
+from llama_stack_api.apis.common.content_types import InterleavedContent
+from llama_stack_api.apis.common.responses import (
     Order,
 )
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.models import Model
-from llama_stack.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1ALPHA
-from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.models import Model
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1ALPHA
+from llama_stack_api.schema_utils import json_schema_type, register_schema, webmethod
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/inspect/inspect.py

@@ -8,11 +8,11 @@ from typing import Literal, Protocol, runtime_checkable
 
 from pydantic import BaseModel
 
-from llama_stack.apis.version import (
+from llama_stack_api.apis.version import (
     LLAMA_STACK_API_V1,
 )
-from llama_stack.providers.datatypes import HealthStatus
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.providers.datatypes import HealthStatus
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 # Valid values for the route filter parameter.
 # Actual API levels: v1, v1alpha, v1beta (filters by level, excludes deprecated)

src/llama-stack-api/llama_stack_api/apis/models/models.py

@@ -9,10 +9,10 @@ from typing import Any, Literal, Protocol, runtime_checkable
 
 from pydantic import BaseModel, ConfigDict, Field, field_validator
 
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.resource import Resource, ResourceType
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.resource import Resource, ResourceType
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 class CommonModelFields(BaseModel):

src/llama-stack-api/llama_stack_api/apis/post_training/post_training.py

@@ -10,11 +10,11 @@ from typing import Annotated, Any, Literal, Protocol
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.common.content_types import URL
-from llama_stack.apis.common.job_types import JobStatus
-from llama_stack.apis.common.training_types import Checkpoint
-from llama_stack.apis.version import LLAMA_STACK_API_V1ALPHA
-from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
+from llama_stack_api.apis.common.content_types import URL
+from llama_stack_api.apis.common.job_types import JobStatus
+from llama_stack_api.apis.common.training_types import Checkpoint
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1ALPHA
+from llama_stack_api.schema_utils import json_schema_type, register_schema, webmethod
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/prompts/prompts.py

@@ -10,9 +10,9 @@ from typing import Protocol, runtime_checkable
 
 from pydantic import BaseModel, Field, field_validator, model_validator
 
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/providers/providers.py

@@ -8,9 +8,9 @@ from typing import Any, Protocol, runtime_checkable
 
 from pydantic import BaseModel
 
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.providers.datatypes import HealthResponse
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.providers.datatypes import HealthResponse
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/safety/safety.py

@@ -9,11 +9,11 @@ from typing import Any, Protocol, runtime_checkable
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.inference import OpenAIMessageParam
-from llama_stack.apis.shields import Shield
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.inference import OpenAIMessageParam
+from llama_stack_api.apis.shields import Shield
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/scoring/scoring.py

@@ -8,9 +8,9 @@ from typing import Any, Protocol, runtime_checkable
 
 from pydantic import BaseModel
 
-from llama_stack.apis.scoring_functions import ScoringFn, ScoringFnParams
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.scoring_functions import ScoringFn, ScoringFnParams
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 # mapping of metric to value
 ScoringResultRow = dict[str, Any]

src/llama-stack-api/llama_stack_api/apis/scoring_functions/scoring_functions.py

@@ -16,10 +16,10 @@ from typing import (
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.common.type_system import ParamType
-from llama_stack.apis.resource import Resource, ResourceType
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
+from llama_stack_api.apis.common.type_system import ParamType
+from llama_stack_api.apis.resource import Resource, ResourceType
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, register_schema, webmethod
 
 
 # Perhaps more structure can be imposed on these functions. Maybe they could be associated

src/llama-stack-api/llama_stack_api/apis/shields/shields.py

@@ -8,10 +8,10 @@ from typing import Any, Literal, Protocol, runtime_checkable
 
 from pydantic import BaseModel
 
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.resource import Resource, ResourceType
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.resource import Resource, ResourceType
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 class CommonShieldFields(BaseModel):

src/llama-stack-api/llama_stack_api/apis/tools/rag_tool.py

@@ -9,7 +9,7 @@ from typing import Annotated, Any, Literal
 
 from pydantic import BaseModel, Field, field_validator
 
-from llama_stack.apis.common.content_types import URL, InterleavedContent
+from llama_stack_api.apis.common.content_types import URL, InterleavedContent
 
 
 class RRFRanker(BaseModel):

src/llama-stack-api/llama_stack_api/apis/tools/tools.py

@@ -10,11 +10,11 @@ from typing import Any, Literal, Protocol
 from pydantic import BaseModel
 from typing_extensions import runtime_checkable
 
-from llama_stack.apis.common.content_types import URL, InterleavedContent
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.resource import Resource, ResourceType
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.apis.common.content_types import URL, InterleavedContent
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.resource import Resource, ResourceType
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/vector_io/vector_io.py

@@ -13,12 +13,12 @@ from typing import Annotated, Any, Literal, Protocol, runtime_checkable
 from fastapi import Body
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.common.tracing import telemetry_traceable
-from llama_stack.apis.inference import InterleavedContent
-from llama_stack.apis.vector_stores import VectorStore
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.schema_utils import json_schema_type, webmethod
-from llama_stack.strong_typing.schema import register_schema
+from llama_stack_api.apis.common.tracing import telemetry_traceable
+from llama_stack_api.apis.inference import InterleavedContent
+from llama_stack_api.apis.vector_stores import VectorStore
+from llama_stack_api.apis.version import LLAMA_STACK_API_V1
+from llama_stack_api.schema_utils import json_schema_type, webmethod
+from llama_stack_api.strong_typing.schema import register_schema
 
 
 @json_schema_type

src/llama-stack-api/llama_stack_api/apis/vector_stores/vector_stores.py

@@ -8,7 +8,7 @@ from typing import Literal
 
 from pydantic import BaseModel
 
-from llama_stack.apis.resource import Resource, ResourceType
+from llama_stack_api.apis.resource import Resource, ResourceType
 
 
 # Internal resource type for storing the vector store routing and other information

src/llama-stack-api/llama_stack_api/providers/__init__.py

@@ -0,0 +1,7 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from .datatypes import *  # noqa: F403

src/llama-stack-api/llama_stack_api/providers/datatypes.py

@@ -10,15 +10,15 @@ from urllib.parse import urlparse
 
 from pydantic import BaseModel, Field
 
-from llama_stack.apis.benchmarks import Benchmark
-from llama_stack.apis.datasets import Dataset
-from llama_stack.apis.datatypes import Api
-from llama_stack.apis.models import Model
-from llama_stack.apis.scoring_functions import ScoringFn
-from llama_stack.apis.shields import Shield
-from llama_stack.apis.tools import ToolGroup
-from llama_stack.apis.vector_stores import VectorStore
-from llama_stack.schema_utils import json_schema_type
+from llama_stack_api.apis.benchmarks import Benchmark
+from llama_stack_api.apis.datasets import Dataset
+from llama_stack_api.apis.datatypes import Api
+from llama_stack_api.apis.models import Model
+from llama_stack_api.apis.scoring_functions import ScoringFn
+from llama_stack_api.apis.shields import Shield
+from llama_stack_api.apis.tools import ToolGroup
+from llama_stack_api.apis.vector_stores import VectorStore
+from llama_stack_api.schema_utils import json_schema_type
 
 
 class ModelsProtocolPrivate(Protocol):

src/llama-stack-api/pyproject.toml

@@ -0,0 +1,82 @@
+[build-system]
+requires = ["setuptools>=61.0"]
+build-backend = "setuptools.build_meta"
+
+[tool.uv]
+required-version = ">=0.7.0"
+
+[project]
+name = "llama-stack-api"
+version = "0.1.0"
+authors = [{ name = "Meta Llama", email = "llama-oss@meta.com" }]
+description = "API and Provider specifications for Llama Stack - lightweight package with protocol definitions and provider specs"
+readme = "README.md"
+requires-python = ">=3.12"
+license = { "text" = "MIT" }
+classifiers = [
+    "License :: OSI Approved :: MIT License",
+    "Programming Language :: Python :: 3",
+    "Operating System :: OS Independent",
+    "Intended Audience :: Developers",
+    "Intended Audience :: Information Technology",
+    "Intended Audience :: Science/Research",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+    "Topic :: Scientific/Engineering :: Information Analysis",
+]
+dependencies = [
+    "pydantic>=2.11.9",
+    "jsonschema",
+    "opentelemetry-sdk>=1.30.0",
+    "opentelemetry-exporter-otlp-proto-http>=1.30.0",
+]
+
+[project.urls]
+Homepage = "https://github.com/llamastack/llama-stack"
+
+[tool.setuptools.packages.find]
+where = ["."]
+include = ["llama_stack_api", "llama_stack_api.*"]
+
+[tool.setuptools.package-data]
+llama_stack_api = ["py.typed"]
+
+[tool.ruff]
+line-length = 120
+
+[tool.ruff.lint]
+select = [
+    "UP",      # pyupgrade
+    "B",       # flake8-bugbear
+    "B9",      # flake8-bugbear subset
+    "C",       # comprehensions
+    "E",       # pycodestyle
+    "F",       # Pyflakes
+    "N",       # Naming
+    "W",       # Warnings
+    "DTZ",     # datetime rules
+    "I",       # isort (imports order)
+    "RUF001",  # Checks for ambiguous Unicode characters in strings
+    "RUF002",  # Checks for ambiguous Unicode characters in docstrings
+    "RUF003",  # Checks for ambiguous Unicode characters in comments
+    "PLC2401", # Checks for the use of non-ASCII characters in variable names
+]
+ignore = [
+    # The following ignores are desired by the project maintainers.
+    "E402",   # Module level import not at top of file
+    "E501",   # Line too long
+    "F405",   # Maybe undefined or defined from star import
+    "C408",   # Ignored because we like the dict keyword argument syntax
+    "N812",   # Ignored because import torch.nn.functional as F is PyTorch convention
+
+    # These are the additional ones we started ignoring after moving to ruff. We should look into each one of them later.
+    "C901",   # Complexity of the function is too high
+]
+unfixable = [
+    "PLE2515",
+] # Do not fix this automatically since ruff will replace the zero-width space with \u200b - let's do it manually
+
+[tool.ruff.lint.per-file-ignores]
+"llama_stack_api/apis/**/__init__.py" = ["F403"]
+
+[tool.ruff.lint.pep8-naming]
+classmethod-decorators = ["classmethod", "pydantic.field_validator"]

src/llama_stack/cli/stack/_list_deps.py

@@ -9,6 +9,7 @@ import sys
 from pathlib import Path
 
 import yaml
+from llama_stack_api.providers.datatypes import Api
 from termcolor import cprint
 
 from llama_stack.cli.stack.utils import ImageType
@@ -21,7 +22,6 @@ from llama_stack.core.datatypes import (
 from llama_stack.core.distribution import get_provider_registry
 from llama_stack.core.stack import replace_env_vars
 from llama_stack.log import get_logger
-from llama_stack.providers.datatypes import Api
 
 TEMPLATES_PATH = Path(__file__).parent.parent.parent / "templates"
 

src/llama_stack/cli/stack/utils.py

@@ -11,6 +11,7 @@ from functools import lru_cache
 from pathlib import Path
 
 import yaml
+from llama_stack_api.providers.datatypes import Api
 from termcolor import cprint
 
 from llama_stack.core.datatypes import (
@@ -32,7 +33,6 @@ from llama_stack.core.storage.datatypes import (
 from llama_stack.core.utils.config_dirs import DISTRIBS_BASE_DIR, EXTERNAL_PROVIDERS_DIR
 from llama_stack.core.utils.dynamic import instantiate_class_type
 from llama_stack.core.utils.image_types import LlamaStackImageType
-from llama_stack.providers.datatypes import Api
 
 TEMPLATES_PATH = Path(__file__).parent.parent.parent / "distributions"
 

src/llama_stack/core/build.py

@@ -7,6 +7,7 @@
 import importlib.resources
 import sys
 
+from llama_stack_api.providers.datatypes import Api
 from pydantic import BaseModel
 from termcolor import cprint
 
@@ -17,7 +18,6 @@ from llama_stack.core.utils.exec import run_command
 from llama_stack.core.utils.image_types import LlamaStackImageType
 from llama_stack.distributions.template import DistributionTemplate
 from llama_stack.log import get_logger
-from llama_stack.providers.datatypes import Api
 
 log = get_logger(name=__name__, category="core")
 

src/llama_stack/core/client.py

@@ -12,11 +12,10 @@ from enum import Enum
 from typing import Any, Union, get_args, get_origin
 
 import httpx
+from llama_stack_api.providers.datatypes import RemoteProviderConfig
 from pydantic import BaseModel, parse_obj_as
 from termcolor import cprint
 
-from llama_stack.providers.datatypes import RemoteProviderConfig
-
 _CLIENT_CLASSES = {}
 
 

src/llama_stack/core/configure.py

@@ -6,6 +6,8 @@
 import textwrap
 from typing import Any
 
+from llama_stack_api.providers.datatypes import Api, ProviderSpec
+
 from llama_stack.core.datatypes import (
     LLAMA_STACK_RUN_CONFIG_VERSION,
     DistributionSpec,
@@ -20,7 +22,6 @@ from llama_stack.core.stack import cast_image_name_to_string, replace_env_vars
 from llama_stack.core.utils.dynamic import instantiate_class_type
 from llama_stack.core.utils.prompt_for_config import prompt_for_config
 from llama_stack.log import get_logger
-from llama_stack.providers.datatypes import Api, ProviderSpec
 
 logger = get_logger(name=__name__, category="core")
 

src/llama_stack/core/conversations/conversations.py

@@ -8,9 +8,7 @@ import secrets
 import time
 from typing import Any, Literal
 
-from pydantic import BaseModel, TypeAdapter
-
-from llama_stack.apis.conversations.conversations import (
+from llama_stack_api.apis.conversations.conversations import (
     Conversation,
     ConversationDeletedResource,
     ConversationItem,
@@ -20,6 +18,8 @@ from llama_stack.apis.conversations.conversations import (
     Conversations,
     Metadata,
 )
+from pydantic import BaseModel, TypeAdapter
+
 from llama_stack.core.datatypes import AccessRule, StackRunConfig
 from llama_stack.log import get_logger
 from llama_stack.providers.utils.sqlstore.api import ColumnDefinition, ColumnType

src/llama_stack/core/datatypes.py

@@ -9,22 +9,23 @@ from pathlib import Path
 from typing import Annotated, Any, Literal, Self
 from urllib.parse import urlparse
 
+from llama_stack_api.apis.benchmarks import Benchmark, BenchmarkInput
+from llama_stack_api.apis.datasetio import DatasetIO
+from llama_stack_api.apis.datasets import Dataset, DatasetInput
+from llama_stack_api.apis.eval import Eval
+from llama_stack_api.apis.inference import Inference
+from llama_stack_api.apis.models import Model, ModelInput
+from llama_stack_api.apis.resource import Resource
+from llama_stack_api.apis.safety import Safety
+from llama_stack_api.apis.scoring import Scoring
+from llama_stack_api.apis.scoring_functions import ScoringFn, ScoringFnInput
+from llama_stack_api.apis.shields import Shield, ShieldInput
+from llama_stack_api.apis.tools import ToolGroup, ToolGroupInput, ToolRuntime
+from llama_stack_api.apis.vector_io import VectorIO
+from llama_stack_api.apis.vector_stores import VectorStore, VectorStoreInput
+from llama_stack_api.providers.datatypes import Api, ProviderSpec
 from pydantic import BaseModel, Field, field_validator, model_validator
 
-from llama_stack.apis.benchmarks import Benchmark, BenchmarkInput
-from llama_stack.apis.datasetio import DatasetIO
-from llama_stack.apis.datasets import Dataset, DatasetInput
-from llama_stack.apis.eval import Eval
-from llama_stack.apis.inference import Inference
-from llama_stack.apis.models import Model, ModelInput
-from llama_stack.apis.resource import Resource
-from llama_stack.apis.safety import Safety
-from llama_stack.apis.scoring import Scoring
-from llama_stack.apis.scoring_functions import ScoringFn, ScoringFnInput
-from llama_stack.apis.shields import Shield, ShieldInput
-from llama_stack.apis.tools import ToolGroup, ToolGroupInput, ToolRuntime
-from llama_stack.apis.vector_io import VectorIO
-from llama_stack.apis.vector_stores import VectorStore, VectorStoreInput
 from llama_stack.core.access_control.datatypes import AccessRule
 from llama_stack.core.storage.datatypes import (
     KVStoreReference,
@@ -32,7 +33,6 @@ from llama_stack.core.storage.datatypes import (
     StorageConfig,
 )
 from llama_stack.log import LoggingConfig
-from llama_stack.providers.datatypes import Api, ProviderSpec
 
 LLAMA_STACK_BUILD_CONFIG_VERSION = 2
 LLAMA_STACK_RUN_CONFIG_VERSION = 2