Cognee API

A FastAPI server wrapper for Cognee that exposes AI memory graph functionality via REST API. Perfect for integration with Claude Code Skills, automation workflows, and building AI-powered applications.

Overview

Cognee API provides a RESTful interface to Cognee's powerful knowledge graph capabilities, enabling you to:

• Ingest and process text data into structured knowledge graphs
• Build semantic relationships between concepts
• Query AI memory using natural language
• Manage and enrich knowledge bases programmatically

Features

• RESTful API - Clean HTTP endpoints for all Cognee operations
• Async by Default - Built on FastAPI for high-performance async operations
• Type Safety - Full Pydantic validation for requests and responses
• Multiple Search Modes - Graph completion, insights extraction, and coding rules
• Multi-tenant Support - User and node set isolation for different contexts
• Production Ready - Health checks, error handling, and structured responses

Quick Start

Installation

pip install fastapi uvicorn cognee httpx

Configuration

Create a .env file based on .env.example:

LLM_API_KEY=your-openai-api-key-here
# LLM_PROVIDER=ollama  # Uncomment for local LLM usage

Running the Server

# Development mode with auto-reload
uvicorn server:app --reload

# Production mode
uvicorn server:app --host 0.0.0.0 --port 8000

The server will start at http://localhost:8000. Visit http://localhost:8000/docs for interactive API documentation.

Deployment to eva_cognee Server

From Development Machine

1. Create your .env file in the project directory with your environment variables:

   LLM_API_KEY=your-api-key-here

2. Run the deployment script:

   ./deploy.sh

Setup on eva_cognee Server

After deploying files to the server, SSH into eva_cognee and set up the environment:

# SSH into the server
ssh root@eva_cognee

# Navigate to the project directory
cd /root/cognee-api

# Create a virtual environment
python3 -m venv venv

# Activate the virtual environment
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Run the server
python server.py

The server will start on http://0.0.0.0:8000 (accessible at http://eva_cognee:8000 or http://192.168.30.33:8000)

Running the Server on eva_cognee

To run the server after initial setup:

cd /root/cognee-api
source venv/bin/activate
python server.py

API Endpoints

Health Check

GET /health

Add Data

POST /add
Content-Type: application/json

{
  "text": "Your text content here",
  "user_id": "optional-user-id",
  "node_set": "optional-node-set",
  "dataset_name": "optional-dataset-name"
}

Build Knowledge Graph

POST /cognify
Content-Type: application/json

{
  "datasets": ["dataset1", "dataset2"]
}

Enrich Graph

POST /memify
Content-Type: application/json

{
  "dataset": "dataset-name",
  "extraction_tasks": [],
  "enrichment_tasks": []
}

Search

POST /search
Content-Type: application/json

{
  "query_text": "What is Cognee?",
  "query_type": "GRAPH_COMPLETION",
  "user_id": "optional-user-id",
  "node_set": "optional-node-set",
  "node_name": []
}

Search Types:

• GRAPH_COMPLETION - General knowledge graph queries
• INSIGHTS - Extract insights and patterns
• CODING_RULES - Query coding-specific rules

Delete Data

POST /delete
Content-Type: application/json

{
  "data_id": "data-id-to-delete"
}

Clear All Data

POST /prune

Using the Python Client

import asyncio
from client import CogneeClient, SearchType

async def main():
    client = CogneeClient(base_url="http://localhost:8000")

    # Ingest data
    add_result = await client.add(
        "Cognee builds AI memory graphs.",
        dataset_name="demo"
    )

    # Build knowledge graph
    await client.cognify(datasets=["demo"])

    # Enrich (optional)
    await client.memify(dataset="demo")

    # Query the graph
    search_result = await client.search(
        "What is Cognee?",
        query_type=SearchType.GRAPH_COMPLETION
    )
    print(search_result["data"])

    await client.close()

asyncio.run(main())

Typical Workflow

1. Ingest - Add your text data using /add
2. Process - Build the knowledge graph with /cognify
3. Enrich - Optionally enhance with /memify
4. Query - Search your knowledge base with /search
5. Manage - Clean up with /delete or /prune

Response Format

All endpoints return consistent JSON responses:

{
  "success": true,
  "data": {},
  "message": "Operation description"
}

Error responses follow the same structure with "success": false.

Development

Project Structure

cognee-api/
├── server.py          # FastAPI server implementation
├── client.py          # Python client library
├── .env.example       # Environment variables template
├── README.md          # This file
└── CLAUDE.md          # Claude Code integration guide

Requirements

• Python 3.8+
• FastAPI
• Uvicorn
• Cognee
• httpx (for client)
• Pydantic

Use Cases

• Claude Code Skills - Build AI assistants with persistent memory
• Documentation Q&A - Ingest docs and query with natural language
• Knowledge Management - Create searchable knowledge bases
• Code Analysis - Extract and query coding patterns and rules
• Research Tools - Build semantic research databases

License

[Add your license here]

Contributing

[Add contribution guidelines here]

Support

For issues and questions:

• Cognee documentation: Cognee GitHub
• FastAPI documentation: FastAPI Docs