feat: Add Google Vertex AI inference provider support

- Add new Vertex AI remote inference provider with litellm integration
- Support for Gemini models through Google Cloud Vertex AI platform
- Uses Google Cloud Application Default Credentials (ADC) for authentication
- Added VertexAI models: gemini-2.5-flash, gemini-2.5-pro, gemini-2.0-flash.
- Updated provider registry to include vertexai provider
- Updated starter template to support Vertex AI configuration
- Added comprehensive documentation and sample configuration

Signed-off-by: Eran Cohen <eranco@redhat.com>

llama_stack/templates/ci-tests/build.yaml

@@ -18,6 +18,7 @@ distribution_spec:
     - remote::openai
     - remote::anthropic
     - remote::gemini
+    - remote::vertexai
     - remote::groq
     - remote::llama-openai-compat
     - remote::sambanova

llama_stack/templates/ci-tests/run.yaml

@@ -85,6 +85,11 @@ providers:
     provider_type: remote::gemini
     config:
       api_key: ${env.GEMINI_API_KEY}
+  - provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+    provider_type: remote::vertexai
+    config:
+      project: ${env.VERTEX_AI_PROJECT}
+      location: ${env.VERTEX_AI_LOCATION:=us-central1}
   - provider_id: ${env.ENABLE_GROQ:=__disabled__}
     provider_type: remote::groq
     config:
@@ -963,6 +968,35 @@ models:
   provider_id: ${env.ENABLE_GEMINI:=__disabled__}
   provider_model_id: gemini/text-embedding-004
   model_type: embedding
+- metadata: {}
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/gemini-2.0-flash
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/gemini-2.0-flash
+  model_type: llm
+- metadata: {}
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/gemini-2.5-flash
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/gemini-2.5-flash
+  model_type: llm
+- metadata: {}
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/gemini-2.5-pro
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/gemini-2.5-pro
+  model_type: llm
+- metadata:
+    embedding_dimension: 768
+    context_length: 2048
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/text-embedding-004
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/text-embedding-004
+  model_type: embedding
+- metadata:
+    embedding_dimension: 768
+    context_length: 2048
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/text-embedding-005
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/text-embedding-005
+  model_type: embedding
 - metadata: {}
   model_id: ${env.ENABLE_GROQ:=__disabled__}/groq/llama3-8b-8192
   provider_id: ${env.ENABLE_GROQ:=__disabled__}

llama_stack/templates/starter/build.yaml

@@ -18,6 +18,7 @@ distribution_spec:
     - remote::openai
     - remote::anthropic
     - remote::gemini
+    - remote::vertexai
     - remote::groq
     - remote::llama-openai-compat
     - remote::sambanova

llama_stack/templates/starter/run.yaml

@@ -85,6 +85,11 @@ providers:
     provider_type: remote::gemini
     config:
       api_key: ${env.GEMINI_API_KEY}
+  - provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+    provider_type: remote::vertexai
+    config:
+      project: ${env.VERTEX_AI_PROJECT}
+      location: ${env.VERTEX_AI_LOCATION:=us-central1}
   - provider_id: ${env.ENABLE_GROQ:=__disabled__}
     provider_type: remote::groq
     config:
@@ -963,6 +968,35 @@ models:
   provider_id: ${env.ENABLE_GEMINI:=__disabled__}
   provider_model_id: gemini/text-embedding-004
   model_type: embedding
+- metadata: {}
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/gemini-2.0-flash
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/gemini-2.0-flash
+  model_type: llm
+- metadata: {}
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/gemini-2.5-flash
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/gemini-2.5-flash
+  model_type: llm
+- metadata: {}
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/gemini-2.5-pro
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/gemini-2.5-pro
+  model_type: llm
+- metadata:
+    embedding_dimension: 768
+    context_length: 2048
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/text-embedding-004
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/text-embedding-004
+  model_type: embedding
+- metadata:
+    embedding_dimension: 768
+    context_length: 2048
+  model_id: ${env.ENABLE_VERTEXAI:=__disabled__}/vertex_ai/text-embedding-005
+  provider_id: ${env.ENABLE_VERTEXAI:=__disabled__}
+  provider_model_id: vertex_ai/text-embedding-005
+  model_type: embedding
 - metadata: {}
   model_id: ${env.ENABLE_GROQ:=__disabled__}/groq/llama3-8b-8192
   provider_id: ${env.ENABLE_GROQ:=__disabled__}

llama_stack/templates/starter/starter.py

@@ -64,6 +64,9 @@ from llama_stack.providers.remote.inference.sambanova.models import (
 from llama_stack.providers.remote.inference.together.models import (
     MODEL_ENTRIES as TOGETHER_MODEL_ENTRIES,
 )
+from llama_stack.providers.remote.inference.vertexai.models import (
+    MODEL_ENTRIES as VERTEXAI_MODEL_ENTRIES,
+)
 from llama_stack.providers.remote.vector_io.chroma.config import ChromaVectorIOConfig
 from llama_stack.providers.remote.vector_io.pgvector.config import (
     PGVectorVectorIOConfig,
@@ -93,6 +96,7 @@ def _get_model_entries_for_provider(provider_type: str) -> list[ProviderModelEnt
         "databricks": DATABRICKS_MODEL_ENTRIES,
         "nvidia": NVIDIA_MODEL_ENTRIES,
         "runpod": RUNPOD_MODEL_ENTRIES,
+        "vertexai": VERTEXAI_MODEL_ENTRIES,
     }
 
     # Special handling for providers with dynamic model entries
@@ -354,6 +358,14 @@ def get_distribution_template() -> DistributionTemplate:
                 "",
                 "Gemini API Key",
             ),
+            "VERTEX_AI_PROJECT": (
+                "",
+                "Google Cloud Project ID for Vertex AI",
+            ),
+            "VERTEX_AI_LOCATION": (
+                "us-central1",
+                "Google Cloud Location for Vertex AI",
+            ),
             "SAMBANOVA_API_KEY": (
                 "",
                 "SambaNova API Key",