This repository is archived. Development has moved to the cognee-integrations monorepo.

• The PyPI package (cognee-integration-claude) is unchanged — pip install cognee-integration-claude still works.
• New releases are published from the monorepo.
• File issues and open PRs at the monorepo.

Cognee-Integration-Claude

A powerful integration between Cognee and Claude Agent SDK that provides intelligent memory management and retrieval capabilities for AI agents.

Overview

cognee-integration-claude combines Cognee's advanced memory layer with Anthropic's Claude Agent SDK. This integration allows you to build AI agents that can efficiently store, search, and retrieve information from a persistent knowledge base.

Features

• Smart Knowledge Storage: Add and persist information using Cognee's advanced indexing
• Semantic Search: Retrieve relevant information using natural language queries
• Session Management: Support for user-specific data isolation
• Claude Agent SDK Integration: Seamless integration with Claude's agent framework
• Async Support: Built with async/await for high-performance applications
• Thread-Safe: Queue-based processing for concurrent operations
• Cross-Session Persistence: Memory survives between agent instances

Installation

pip install cognee-integration-claude

Or using uv:

uv add cognee-integration-claude

Quick Start

import asyncio
import os
from dotenv import load_dotenv
import cognee
from claude_agent_sdk import (
    create_sdk_mcp_server,
    ClaudeAgentOptions,
    ClaudeSDKClient,
    AssistantMessage,
    TextBlock,
)
from cognee_integration_claude import add_tool, search_tool

load_dotenv()

async def main():
    # Clean up memory to start fresh (Optional)
    await cognee.prune.prune_data()
    await cognee.prune.prune_system(metadata=True)
    
    # Create an MCP server with Cognee tools
    server = create_sdk_mcp_server(
        name="cognee-tools",
        version="1.0.0",
        tools=[add_tool, search_tool]
    )
    
    # Configure the agent
    options = ClaudeAgentOptions(
        mcp_servers={"tools": server},
        allowed_tools=["mcp__tools__add_tool", "mcp__tools__search_tool"],
    )
    
    # Use the agent to store information
    async with ClaudeSDKClient(options=options) as client:
        await client.query(
            "Remember that our company signed a contract with HealthBridge Systems "
            "in the healthcare industry, starting Feb 2023, ending Jan 2026, worth £2.4M"
        )
        
        async for msg in client.receive_response():
            if isinstance(msg, AssistantMessage):
                for block in msg.content:
                    if isinstance(block, TextBlock):
                        print(f"Claude: {block.text}")
    
    # Query the stored information (new agent instance)
    async with ClaudeSDKClient(options=options) as client:
        await client.query("What contracts do we have in the healthcare industry?")
        
        async for msg in client.receive_response():
            if isinstance(msg, AssistantMessage):
                for block in msg.content:
                    if isinstance(block, TextBlock):
                        print(f"Claude: {block.text}")

if __name__ == "__main__":
    asyncio.run(main())

Available Tools

Basic Tools

from cognee_integration_claude import add_tool, search_tool

# add_tool: Store information in the knowledge base
# search_tool: Search and retrieve previously stored information

Sessionized Tools

For multi-user applications, use sessionized tools to isolate data between users:

from cognee_integration_claude import get_sessionized_cognee_tools

# Get tools for a specific user session
add_tool, search_tool = get_sessionized_cognee_tools("user-123")

# Auto-generate a session ID
add_tool, search_tool = get_sessionized_cognee_tools()

Session Management

cognee-integration-claude supports user-specific sessions to tag data and isolate retrieval between different users or contexts:

import asyncio
from claude_agent_sdk import (
    create_sdk_mcp_server,
    ClaudeAgentOptions,
    ClaudeSDKClient,
    AssistantMessage,
    TextBlock,
)
from cognee_integration_claude import get_sessionized_cognee_tools

async def main():
    # Each user gets their own isolated session
    user1_add, user1_search = get_sessionized_cognee_tools("user-123")
    user2_add, user2_search = get_sessionized_cognee_tools("user-456")
    
    # Create separate agent configurations for each user
    user1_server = create_sdk_mcp_server(
        name="user1-tools",
        version="1.0.0",
        tools=[user1_add, user1_search]
    )
    
    user2_server = create_sdk_mcp_server(
        name="user2-tools",
        version="1.0.0",
        tools=[user2_add, user2_search]
    )
    
    user1_options = ClaudeAgentOptions(
        mcp_servers={"tools": user1_server},
        allowed_tools=["mcp__tools__add_tool", "mcp__tools__search_tool"],
    )
    
    user2_options = ClaudeAgentOptions(
        mcp_servers={"tools": user2_server},
        allowed_tools=["mcp__tools__add_tool", "mcp__tools__search_tool"],
    )
    
    # Each agent works with isolated data
    async with ClaudeSDKClient(options=user1_options) as client:
        await client.query("Remember: I like pizza")
        async for msg in client.receive_response():
            pass
    
    async with ClaudeSDKClient(options=user2_options) as client:
        await client.query("Remember: I like sushi")
        async for msg in client.receive_response():
            pass
    
    # User 1 can only see their own data
    async with ClaudeSDKClient(options=user1_options) as client:
        await client.query("What food do I like?")
        async for msg in client.receive_response():
            if isinstance(msg, AssistantMessage):
                for block in msg.content:
                    if isinstance(block, TextBlock):
                        print(f"User 1's agent: {block.text}")  # Will mention pizza, not sushi

if __name__ == "__main__":
    asyncio.run(main())

Tool Reference

add_tool(data: str)

Store information in the memory for later retrieval.

Parameters:

• data (str): The text or information you want to store

Returns: Confirmation message

Example:

server = create_sdk_mcp_server(
    name="memory-tools",
    version="1.0.0",
    tools=[add_tool]
)

options = ClaudeAgentOptions(
    mcp_servers={"tools": server},
    allowed_tools=["mcp__tools__add_tool"],
)

async with ClaudeSDKClient(options=options) as client:
    await client.query("Store this: Our Q4 revenue was $2.5M with 15% growth")
    async for msg in client.receive_response():
        pass

search_tool(query_text: str)

Search and retrieve previously stored information from the memory.

Parameters:

• query_text (str): Natural language search query

Returns: List of relevant search results

Example:

server = create_sdk_mcp_server(
    name="memory-tools",
    version="1.0.0",
    tools=[search_tool]
)

options = ClaudeAgentOptions(
    mcp_servers={"tools": server},
    allowed_tools=["mcp__tools__search_tool"],
)

async with ClaudeSDKClient(options=options) as client:
    await client.query("What was our Q4 revenue?")
    async for msg in client.receive_response():
        if isinstance(msg, AssistantMessage):
            for block in msg.content:
                if isinstance(block, TextBlock):
                    print(block.text)

get_sessionized_cognee_tools(session_id: Optional[str] = None)

Returns cognee tools with optional user-specific sessionization.

Parameters:

• session_id (Optional[str]): User identifier for data isolation. If not provided, a random session ID is auto-generated.

Returns: [add_tool, search_tool] - A list of sessionized tools

Example:

# With explicit session ID
add_tool, search_tool = get_sessionized_cognee_tools("user-123")

# Auto-generate session ID
add_tool, search_tool = get_sessionized_cognee_tools()

Configuration

Environment Variables

Create a .env file in your project root:

# OpenAI API key (used by Cognee for LLM operations)
LLM_API_KEY=your-openai-api-key-here

You can also use other LLM providers with Cognee. Check the Cognee documentation for more details.

The Claude Agent SDK uses a bundled Claude Code CLI that handles authentication automatically.

1. If you're using Cursor: You're likely already authenticated through your Cursor/Claude session. The bundled CLI will use your existing credentials.

2. If you're running standalone: The first time you run the SDK, it will guide you through authentication via OAuth or API key through the bundled CLI.

3. For CI/CD or automated environments: You may need to authenticate the CLI separately. See the Claude Agent SDK documentation for details.

Cognee Configuration (Optional)

You can customize Cognee's data and system directories:

from cognee.api.v1.config import config
import os

config.data_root_directory(
    os.path.join(os.path.dirname(__file__), ".cognee/data_storage")
)

config.system_root_directory(
    os.path.join(os.path.dirname(__file__), ".cognee/system")
)

Examples

Check out the examples/ directory for simple examples:

• examples/example.py: Basic usage with add and search tools

And the interactive guide:

• cognee_integration_claude/guide.ipynb: Step-by-step Jupyter notebook tutorial

Advanced Usage

Pre-loading Data

You can pre-load data into Cognee before creating agents:

import asyncio
import cognee
from claude_agent_sdk import (
    create_sdk_mcp_server,
    ClaudeAgentOptions,
    ClaudeSDKClient,
    AssistantMessage,
    TextBlock,
)
from cognee_integration_claude import search_tool

async def main():
    # Pre-load data directly into Cognee
    await cognee.add("Important company information here...")
    await cognee.add("More data to remember...")
    await cognee.cognify()  # Process and index the data
    
    # Now create an agent that can search this data
    server = create_sdk_mcp_server(
        name="search-tools",
        version="1.0.0",
        tools=[search_tool]
    )
    
    options = ClaudeAgentOptions(
        mcp_servers={"tools": server},
        allowed_tools=["mcp__tools__search_tool"],
    )
    
    async with ClaudeSDKClient(options=options) as client:
        await client.query("What information do we have?")
        
        async for msg in client.receive_response():
            if isinstance(msg, AssistantMessage):
                for block in msg.content:
                    if isinstance(block, TextBlock):
                        print(block.text)

if __name__ == "__main__":
    asyncio.run(main())

Data Management

import asyncio
import cognee

async def reset_knowledge_base():
    """Clear all data and reset the knowledge base"""
    await cognee.prune.prune_data()
    await cognee.prune.prune_system(metadata=True)

async def visualize_knowledge_graph():
    """Generate a visualization of the knowledge graph"""
    await cognee.visualize_graph("graph.html")

Disabling Default Cursor Tools

When using Claude Agent SDK in environments like Cursor, you may want to disable default tools:

options = ClaudeAgentOptions(
    mcp_servers={"tools": server},
    allowed_tools=["mcp__tools__add_tool", "mcp__tools__search_tool"],
    disallowed_tools=[
        "Task", "Bash", "Glob", "Grep", "ExitPlanMode",
        "Read", "Edit", "Write", "NotebookEdit", "WebFetch",
        "TodoWrite", "WebSearch", "BashOutput", "KillShell", "SlashCommand",
    ],
)

Requirements

• Python 3.13+
• OpenAI API key (or other LLM provider supported by Cognee)

Related Projects

• cognee - The core Cognee memory layer
• claude-agent-sdk - Claude Agent SDK

License

MIT License

Contributing

Contributions are welcome! Please feel free fork to submit a PR.