LightRAG Vector Database

Setup

uv pip install -U agno openai

LightRag connects to a running LightRAG server over HTTP. The default server URL is http://localhost:9621.

export LIGHTRAG_SERVER_URL="http://localhost:9621"
export LIGHTRAG_API_KEY="your-api-key"  # if your server requires one

Example

agent_with_knowledge.py

from os import getenv

from agno.agent import Agent
from agno.knowledge.knowledge import Knowledge
from agno.vectordb.lightrag import LightRag

vector_db = LightRag(
    server_url=getenv("LIGHTRAG_SERVER_URL", "http://localhost:9621"),
    api_key=getenv("LIGHTRAG_API_KEY"),
)

knowledge = Knowledge(
    name="LightRAG Knowledge Base",
    vector_db=vector_db,
)

knowledge.insert(
    name="CV",
    path="data/cv_1.pdf",
)

agent = Agent(
    knowledge=knowledge,
    search_knowledge=True,
)
agent.print_response("What skills does Jordan Mitchell have?", markdown=True)

Inserted content is uploaded to the LightRAG server for indexing. Searches query the server in hybrid mode and return its response as a document, with source references preserved in meta_data["references"].

Search filters are not supported. Indexing happens on the server after upload, so documents may take a moment to become searchable.

Async Support ⚡

LightRAG also supports asynchronous operations, enabling concurrency and leading to better performance.

async_lightrag_db.py

import asyncio
from os import getenv

from agno.agent import Agent
from agno.knowledge.knowledge import Knowledge
from agno.vectordb.lightrag import LightRag

vector_db = LightRag(
    server_url=getenv("LIGHTRAG_SERVER_URL", "http://localhost:9621"),
    api_key=getenv("LIGHTRAG_API_KEY"),
)

knowledge = Knowledge(
    name="LightRAG Knowledge Base",
    vector_db=vector_db,
)

agent = Agent(
    knowledge=knowledge,
    search_knowledge=True,
)

if __name__ == "__main__":
    asyncio.run(
        knowledge.ainsert(
            name="CV",
            path="data/cv_1.pdf",
        )
    )

    asyncio.run(
        agent.aprint_response("What skills does Jordan Mitchell have?", markdown=True)
    )

Use ainsert() and aprint_response() methods with asyncio.run() for non-blocking operations in high-throughput applications.

LightRag Params

Parameter	Type	Default	Description
`server_url`	`str`	`"http://localhost:9621"`	URL of the LightRAG server.
`api_key`	`str`	`None`	API key for the LightRAG server.
`auth_header_name`	`str`	`"X-API-KEY"`	Name of the authentication header.
`auth_header_format`	`str`	`"{api_key}"`	Format string for the authentication header value.
`name`	`str`	`None`	Name of the vector database.
`description`	`str`	`None`	Description of the vector database.

Get Started

Basics

Advanced

Production

Providers

Other

Additional Resources

LightRAG Vector Database

Setup

Example

Async Support ⚡

LightRag Params

​Setup

​Example

Async Support ⚡

​LightRag Params

Setup

Example

LightRag Params