Skip to main content

Using Vector Stores (Knowledge Bases)

Use Vector Stores with any LiteLLM supported model

LiteLLM integrates with vector stores, allowing your models to access your organization's data for more accurate and contextually relevant responses.

Supported Vector Stores​

Quick Start​

To use a vector store with LiteLLM, you need to

  • Initialize litellm.vector_store_registry
  • Pass tools with vector_store_ids to the completion request. Where vector_store_ids is a list of vector store ids you initialized in litellm.vector_store_registry

LiteLLM Python SDK​

LiteLLM's allows you to use vector stores in the OpenAI API spec by passing a tool with vector_store_ids you want to use

Basic Bedrock Knowledge Base Usage
import os
import litellm

from litellm.vector_stores.vector_store_registry import VectorStoreRegistry, LiteLLM_ManagedVectorStore

# Init vector store registry
litellm.vector_store_registry = VectorStoreRegistry(
vector_stores=[
LiteLLM_ManagedVectorStore(
vector_store_id="T37J8R4WTM",
custom_llm_provider="bedrock"
)
]
)


# Make a completion request with vector_store_ids parameter
response = await litellm.acompletion(
model="anthropic/claude-sonnet-5",
messages=[{"role": "user", "content": "What is litellm?"}],
tools=[
{
"type": "file_search",
"vector_store_ids": ["T37J8R4WTM"]
}
],
)

print(response.choices[0].message.content)

LiteLLM Proxy​

1. Configure your vector_store_registry​

To use a vector store with LiteLLM, you need to configure your vector_store_registry. This tells litellm which vector stores to use and api provider to use for the vector store.

config.yaml
model_list:
- model_name: claude-sonnet-5
litellm_params:
model: anthropic/claude-sonnet-5
api_key: os.environ/ANTHROPIC_API_KEY

vector_store_registry:
- vector_store_name: "bedrock-litellm-website-knowledgebase"
litellm_params:
vector_store_id: "T37J8R4WTM"
custom_llm_provider: "bedrock"
vector_store_description: "Bedrock vector store for the Litellm website knowledgebase"
vector_store_metadata:
source: "https://www.litellm.com/docs"

2. Make a request with vector_store_ids parameter​

Curl Request to LiteLLM Proxy
curl http://localhost:4000/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $LITELLM_API_KEY" \
-d '{
"model": "claude-sonnet-5",
"messages": [{"role": "user", "content": "What is litellm?"}],
"tools": [
{
"type": "file_search",
"vector_store_ids": ["T37J8R4WTM"]
}
]
}'

Provider Specific Guides​

This section covers how to add your vector stores to LiteLLM. If you want support for a new provider, please file an issue here.

Bedrock Knowledge Bases​

1. Set up your Bedrock Knowledge Base

Ensure you have a Bedrock Knowledge Base created in your AWS account with the appropriate permissions configured.

2. Add to LiteLLM UI

  1. Navigate to Tools > Vector Stores > "Add new vector store"
  2. Select "Bedrock" as the provider
  3. Enter your Bedrock Knowledge Base ID in the "Vector Store ID" field

Vertex AI RAG Engine​

1. Get your Vertex AI RAG Engine ID

  1. Navigate to your RAG Engine Corpus in the Google Cloud Console
  2. Select the RAG Engine you want to integrate with LiteLLM
  1. Click the "Details" button and copy the UUID for the RAG Engine
  2. The ID should look like: 6917529027641081856

2. Add to LiteLLM UI

  1. Navigate to Tools > Vector Stores > "Add new vector store"
  2. Select "Vertex AI RAG Engine" as the provider
  3. Enter your Vertex AI RAG Engine ID in the "Vector Store ID" field

PG Vector​

1. Deploy the litellm-pg-vector-store connector

LiteLLM provides a server that exposes OpenAI-compatible vector_store endpoints for PG Vector. The LiteLLM Proxy server connects to your deployed service and uses it as a vector store when querying.

  1. Follow the deployment instructions for the litellm-pg-vector-store connector here
  2. For detailed configuration options, see the configuration guide

Example .env configuration for deploying litellm-pg-vector-store:

DATABASE_URL="postgresql://neondb_owner:xxxx"
SERVER_API_KEY="sk-<your-litellm-api-key>"
HOST="0.0.0.0"
PORT=8001
EMBEDDING__MODEL="text-embedding-ada-002"
EMBEDDING__BASE_URL="http://localhost:4000"
EMBEDDING__API_KEY="sk-<your-litellm-api-key>"
EMBEDDING__DIMENSIONS=1536
DB_FIELDS__ID_FIELD="id"
DB_FIELDS__CONTENT_FIELD="content"
DB_FIELDS__METADATA_FIELD="metadata"
DB_FIELDS__EMBEDDING_FIELD="embedding"
DB_FIELDS__VECTOR_STORE_ID_FIELD="vector_store_id"
DB_FIELDS__CREATED_AT_FIELD="created_at"

2. Add to LiteLLM UI

Once your litellm-pg-vector-store is deployed:

  1. Navigate to Tools > Vector Stores > "Add new vector store"
  2. Select "PG Vector" as the provider
  3. Enter your API Base URL and API Key for your litellm-pg-vector-store container
    • The API Key field corresponds to the SERVER_API_KEY from your .env configuration

OpenAI Vector Stores​

1. Set up your OpenAI Vector Store

  1. Create your Vector Store on the OpenAI platform
  2. Note your Vector Store ID (format: vs_687ae3b2439881918b433cb99d10662e)

2. Add to LiteLLM UI

  1. Navigate to Tools > Vector Stores > "Add new vector store"
  2. Select "OpenAI" as the provider
  3. Enter your Vector Store ID in the corresponding field
  4. Enter your OpenAI API Key in the API Key field

Define Vector Store Fallbacks​

Configure an ordered list of backup vector stores for proxy search requests. LiteLLM tries the primary first, then the configured backups when a search fails with a retryable error. Callers keep using the primary store ID or the existing passthrough URL.

Config only

Define the primary and every backup in config.yaml under vector_store_registry. Fallbacks cannot currently be configured through the management API or Admin UI, and database-only stores cannot be used as fallback targets. Database-backed stores still support ordinary searches.

Configure the primary and backups​

Add fallbacks under the primary store's litellm_params. Each entry is another registration's vector_store_id, not its vector_store_name. Use the existing provider store IDs in this OpenAI example:

config.yaml
vector_store_registry:
- vector_store_name: documents-primary
litellm_params:
vector_store_id: vs_primary
custom_llm_provider: openai
api_key: os.environ/PRIMARY_OPENAI_API_KEY
fallbacks: [vs_backup]

- vector_store_name: documents-backup
litellm_params:
vector_store_id: vs_backup
custom_llm_provider: openai
api_key: os.environ/BACKUP_OPENAI_API_KEY

router_settings:
cooldown_time: 5

Replace vs_primary and vs_backup with your real store IDs. Deploy or reload the proxy configuration after changes. Each attempt uses the target store's own provider settings and credentials. The caller must have access to the primary and all configured backups; fallback routing does not bypass store or index permissions.

Backup IDs must be unique and refer to existing YAML registrations that support search. A store cannot fall back to itself, and a backup cannot define another fallback chain. Keep equivalent data in the stores yourself. LiteLLM does not copy documents, replicate writes, or synchronize indexes.

Search through the unified API​

curl -i 'http://localhost:4000/v1/vector_stores/vs_primary/search' \
-H "Authorization: Bearer $LITELLM_API_KEY" \
-H 'Content-Type: application/json' \
-d '{"query": "What is our refund policy?"}'

The unified search API can fall back across providers, provided each store supports search and has compatible data and search settings. These fallbacks apply to the proxy search endpoint and the native routes below; they do not configure chat-completion model fallbacks or provider-managed file-search tools.

Search through native passthrough APIs​

Native search uses the same litellm_params.fallbacks configuration. All targets must use the same provider and API type so the original request and response formats remain valid.

ProviderNative requests that support fallback
Azure AI SearchPOST /azure_ai/indexes/{name}/docs/search, GET /azure_ai/indexes/{name}/docs. The /azure/ prefix is also supported.
MilvusPOST /milvus/v2/vectordb/entities/search
OpenSearchGET or POST /opensearch/{name}/_search
Vertex AI SearchPOST /vertex_ai/discovery/v1*/projects/{project}/locations/{location}/collections/{collection}/dataStores/{store}/servingConfigs/{config}:search

For Azure, Milvus, and OpenSearch, retain the existing index registration pointing to the primary store. You do not need a second LiteLLM index registration for the backup. The backup must contain the same physical index or collection name used by the native request. Vertex supports vertex_ai/search_api datastore searches, not engine routes, and uses each target's project and datastore settings. Milvus hybrid search and other native reads are outside this fallback support.

Writes do not fail over. Document uploads, inserts, updates, deletes, and index lifecycle operations keep their existing passthrough routing. A primary write failure does not send the write to a backup, even when searches use that backup during cooldown.

Failover, cooldown, and response headers​

HTTP 408, 429, and 5xx responses, plus connection failures and timeouts, trigger search failover. Other 4xx responses, including authentication, authorization, and missing-index errors, are terminal. Query-embedding failures do not trigger vector-store failover; configure embedding resiliency on the embedding model group.

A retryable store failure immediately puts that store in cooldown for router_settings.cooldown_time seconds (5 by default). Searches skip it during cooldown. Once the time expires, it becomes eligible again and live traffic tests it. This does not use model-routing failure-count thresholds. If every target is in cooldown, LiteLLM returns 503 with Retry-After.

For cooldown shared across proxy workers and instances, point general_settings.coordination_redis at the same Redis instance in every proxy's configuration:

general_settings:
coordination_redis:
host: os.environ/REDIS_HOST
port: os.environ/REDIS_PORT
password: os.environ/REDIS_PASSWORD

Without coordination Redis, or if it is unavailable, cooldown relies on each worker's local cache.

Successful responses expose x-litellm-vector-store-id for the serving store and x-litellm-attempted-fallbacks for the number of backup attempts. Native responses keep the provider's response body. These headers do not provide a full attempt history.

Advanced​

Logging Vector Store Usage​

LiteLLM allows you to view your vector store usage in the LiteLLM UI on the Logs page.

After completing a request with a vector store, navigate to the Logs page on LiteLLM. Here you should be able to see the query sent to the vector store and corresponding response with scores.

LiteLLM Logs Page: Vector Store Usage

Listing available vector stores​

You can list all available vector stores using the /vector_store/list endpoint

Request:

List all available vector stores
curl -X GET "http://localhost:4000/vector_store/list" \
-H "Authorization: Bearer $LITELLM_API_KEY"

Response:

The response will be a list of all vector stores that are available to use with LiteLLM.

{
"object": "list",
"data": [
{
"vector_store_id": "T37J8R4WTM",
"custom_llm_provider": "bedrock",
"vector_store_name": "bedrock-litellm-website-knowledgebase",
"vector_store_description": "Bedrock vector store for the Litellm website knowledgebase",
"vector_store_metadata": {
"source": "https://www.litellm.com/docs"
},
"created_at": "2023-05-03T18:21:36.462Z",
"updated_at": "2023-05-03T18:21:36.462Z",
"litellm_credential_name": "bedrock_credentials"
}
],
"total_count": 1,
"current_page": 1,
"total_pages": 1
}

Always on for a model​

Use this if you want vector stores to be used by default for a specific model.

In this config, we add vector_store_ids to the claude-3-5-sonnet-with-vector-store model. This means that any request to the claude-3-5-sonnet-with-vector-store model will always use the vector store with the id T37J8R4WTM defined in the vector_store_registry.

Always on for a model
model_list:
- model_name: claude-3-5-sonnet-with-vector-store
litellm_params:
model: anthropic/claude-sonnet-5
vector_store_ids: ["T37J8R4WTM"]

vector_store_registry:
- vector_store_name: "bedrock-litellm-website-knowledgebase"
litellm_params:
vector_store_id: "T37J8R4WTM"
custom_llm_provider: "bedrock"
vector_store_description: "Bedrock vector store for the Litellm website knowledgebase"
vector_store_metadata:
source: "https://www.litellm.com/docs"

How It Works​

If your request includes a vector_store_ids parameter where any of the vector store ids are found in the vector_store_registry, LiteLLM will automatically use the vector store for the request.

  1. You make a completion request with the vector_store_ids parameter and any of the vector store ids are found in the litellm.vector_store_registry
  2. LiteLLM automatically:
    • Uses your last message as the query to retrieve relevant information from the Knowledge Base
    • Adds the retrieved context to your conversation
    • Sends the augmented messages to the model

Example Transformation​

When you pass vector_store_ids=["YOUR_KNOWLEDGE_BASE_ID"], your request flows through these steps:

1. Original Request to LiteLLM:

{
"model": "anthropic/claude-sonnet-5",
"messages": [
{"role": "user", "content": "What is litellm?"}
],
"vector_store_ids": ["YOUR_KNOWLEDGE_BASE_ID"]
}

2. Request to AWS Bedrock Knowledge Base:

{
"retrievalQuery": {
"text": "What is litellm?"
}
}

This is sent to: https://bedrock-agent-runtime.{aws_region}.amazonaws.com/knowledgebases/YOUR_KNOWLEDGE_BASE_ID/retrieve

3. Final Request to LiteLLM:

{
"model": "anthropic/claude-sonnet-5",
"messages": [
{"role": "user", "content": "What is litellm?"},
{"role": "user", "content": "Context: \n\nLiteLLM is an open-source SDK to simplify LLM API calls across providers (OpenAI, Claude, etc). It provides a standardized interface with robust error handling, streaming, and observability tools."}
]
}

This process happens automatically whenever you include the vector_store_ids parameter in your request.

Accessing Search Results (Citations)​

When using vector stores, LiteLLM automatically returns search results in provider_specific_fields. This allows you to show users citations for the AI's response.

Key Concept​

Search results are always in: response.choices[0].message.provider_specific_fields["search_results"]

For streaming: Results appear in the final chunk when finish_reason == "stop"

Non-Streaming Example​

Non-Streaming Response with search results:

{
"id": "chatcmpl-abc123",
"choices": [{
"index": 0,
"message": {
"role": "assistant",
"content": "LiteLLM is a platform...",
"provider_specific_fields": {
"search_results": [{
"search_query": "What is litellm?",
"data": [{
"score": 0.95,
"content": [{"text": "...", "type": "text"}],
"filename": "litellm-docs.md",
"file_id": "doc-123"
}]
}]
}
},
"finish_reason": "stop"
}]
}
from openai import OpenAI

client = OpenAI(
base_url="http://localhost:4000",
api_key="your-litellm-api-key"
)

response = client.chat.completions.create(
model="claude-sonnet-5",
messages=[{"role": "user", "content": "What is litellm?"}],
tools=[{"type": "file_search", "vector_store_ids": ["T37J8R4WTM"]}]
)

# Get AI response
print(response.choices[0].message.content)

# Get search results (citations)
search_results = response.choices[0].message.provider_specific_fields.get("search_results", [])

for result_page in search_results:
for idx, item in enumerate(result_page['data'], 1):
print(f"[{idx}] {item.get('filename', 'Unknown')} (score: {item['score']:.2f})")

Streaming Example​

Streaming Response with search results (final chunk):

{
"id": "chatcmpl-abc123",
"choices": [{
"index": 0,
"delta": {
"provider_specific_fields": {
"search_results": [{
"search_query": "What is litellm?",
"data": [{
"score": 0.95,
"content": [{"text": "...", "type": "text"}],
"filename": "litellm-docs.md",
"file_id": "doc-123"
}]
}]
}
},
"finish_reason": "stop"
}]
}
from openai import OpenAI

client = OpenAI(
base_url="http://localhost:4000",
api_key="your-litellm-api-key"
)

stream = client.chat.completions.create(
model="claude-sonnet-5",
messages=[{"role": "user", "content": "What is litellm?"}],
tools=[{"type": "file_search", "vector_store_ids": ["T37J8R4WTM"]}],
stream=True
)

for chunk in stream:
# Stream content
if chunk.choices[0].delta.content:
print(chunk.choices[0].delta.content, end="", flush=True)

# Get citations in final chunk
if chunk.choices[0].finish_reason == "stop":
search_results = getattr(chunk.choices[0].delta, 'provider_specific_fields', {}).get('search_results', [])
if search_results:
print("\n\nSources:")
for page in search_results:
for idx, item in enumerate(page['data'], 1):
print(f" [{idx}] {item.get('filename', 'Unknown')} ({item['score']:.2f})")

Search Result Fields​

FieldTypeDescription
search_querystringThe query used to search the vector store
dataarrayArray of search results
data[].scorefloatRelevance score (0-1, higher is more relevant)
data[].contentarrayContent chunks with text and type
data[].filenamestringName of the source file (optional)
data[].file_idstringIdentifier for the source file (optional)
data[].attributesobjectProvider-specific metadata (optional)

API Reference​

LiteLLM Completion Knowledge Base Parameters​

When using the Knowledge Base integration with LiteLLM, you can include the following parameters:

ParameterTypeDescription
vector_store_idsList[str]List of Knowledge Base IDs to query

VectorStoreRegistry​

The VectorStoreRegistry is a central component for managing vector stores in LiteLLM. It acts as a registry where you can configure and access your vector stores.

What is VectorStoreRegistry?​

VectorStoreRegistry is a class that:

  • Maintains a collection of vector stores that LiteLLM can use
  • Allows you to register vector stores with their credentials and metadata
  • Makes vector stores accessible via their IDs in your completion requests

Using VectorStoreRegistry in Python​

from litellm.vector_stores.vector_store_registry import VectorStoreRegistry, LiteLLM_ManagedVectorStore

# Initialize the vector store registry with one or more vector stores
litellm.vector_store_registry = VectorStoreRegistry(
vector_stores=[
LiteLLM_ManagedVectorStore(
vector_store_id="YOUR_VECTOR_STORE_ID", # Required: Unique ID for referencing this store
custom_llm_provider="bedrock" # Required: Provider (e.g., "bedrock")
)
]
)

LiteLLM_ManagedVectorStore Parameters​

Each vector store in the registry is configured using a LiteLLM_ManagedVectorStore object with these parameters:

ParameterTypeRequiredDescription
vector_store_idstrYesUnique identifier for the vector store
custom_llm_providerstrYesThe provider of the vector store (e.g., "bedrock")
vector_store_namestrNoA friendly name for the vector store
vector_store_descriptionstrNoDescription of what the vector store contains
vector_store_metadatadict or strNoAdditional metadata about the vector store
litellm_credential_namestrNoName of the credentials to use for this vector store

Configuring VectorStoreRegistry in config.yaml​

For the LiteLLM Proxy, you can configure the same registry in your config.yaml file:

Vector store configuration in config.yaml
vector_store_registry:
- vector_store_name: "bedrock-litellm-website-knowledgebase" # Optional friendly name
litellm_params:
vector_store_id: "T37J8R4WTM" # Required: Unique ID
custom_llm_provider: "bedrock" # Required: Provider
vector_store_description: "Bedrock vector store for the Litellm website knowledgebase"
vector_store_metadata:
source: "https://www.litellm.com/docs"

The litellm_params section accepts all the same parameters as the LiteLLM_ManagedVectorStore constructor in the Python SDK.