Cognee

What is Cognee about?

Cognee provides a modular, scalable pipeline (Extract, Cognify, Load) that turns raw data into a knowledge graph and vector store, allowing agents to query their own memory instead of relying on traditional Retrieval‑Augmented Generation (RAG) setups.

How to use Cognee?

1. Install the package:

   pip install cognee
   

2. Set your LLM API key (e.g., OpenAI) via environment variable or a .env file.
3. Add data, run the cognition pipeline, and query:

   import cognee, asyncio, os
   os.environ["LLM_API_KEY"] = "YOUR_OPENAI_API_KEY"
   
   async def main():
       await cognee.add("Natural language processing (NLP) …")
       await cognee.cognify()
       results = await cognee.search("Tell me about NLP")
       for r in results:
           print(r)
   
   asyncio.run(main())
   

4. Optionally launch the UI for visual interaction via the documentation link.

Key Features of Cognee

• One‑line data ingestion using Pydantic models.
• Automatic extraction, knowledge graph creation, and vector embedding.
• Supports 30+ data sources (text, PDFs, images, audio transcriptions, etc.).
• Works with Neo4j, other graph DBs, and vector stores.
• Replaces conventional RAG pipelines, reducing developer effort and cost.
• Fully hosted beta (Cogwit) and open‑source UI for local exploration.

Use Cases of Cognee

• Building dynamic, long‑term memory for autonomous agents.
• Enhancing chatbot context retention across sessions.
• Knowledge management for enterprises (documents, manuals, media).
• Rapid prototyping of GraphRAG applications.
• Academic research requiring searchable corpora linked to semantics.

FAQ

Q: Which Python versions are supported? A: 3.10 – 3.13.

Q: Do I need a specific database? A: Cognee can work with any compatible graph or vector database; Neo4j is the primary example.

Q: Can I run Cognee locally without an LLM provider? A: Yes, by configuring an open‑source LLM endpoint (e.g., Ollama) and setting the appropriate environment variables.

Q: How can I contribute? A: See CONTRIBUTING.md in the repo; the project welcomes first‑time contributors and sponsors.

Q: Is there a hosted service? A: Cogwit beta offers a fully hosted AI memory platform; sign‑up at https://platform.cognee.ai/.

cognee - Memory for AI Agents in 5 lines of code

🚀 We launched Cogwit beta (Fully-hosted AI Memory): Sign up here! 🚀

Build dynamic memory for Agents and replace RAG using scalable, modular ECL (Extract, Cognify, Load) pipelines.

More on use-cases and evals

Features

• Interconnect and retrieve your past conversations, documents, images and audio transcriptions
• Replaces RAG systems and reduces developer effort, and cost.
• Load data to graph and vector databases using only Pydantic
• Manipulate your data while ingesting from 30+ data sources

Get Started

Get started quickly with a Google Colab notebook , Deepnote notebook or starter repo

Contributing

Your contributions are at the core of making this a true open source project. Any contributions you make are greatly appreciated. See CONTRIBUTING.md for more information.

📦 Installation

You can install Cognee using either pip, poetry, uv or any other python package manager. Cognee supports Python 3.10 to 3.13

With pip

pip install cognee

Local Cognee installation

You can install the local Cognee repo using pip, poetry and uv. For local pip installation please make sure your pip version is above version 21.3.

with UV with all optional dependencies

uv sync --all-extras

💻 Basic Usage

Setup

import os
os.environ["LLM_API_KEY"] = "YOUR OPENAI_API_KEY"

You can also set the variables by creating .env file, using our template. To use different LLM providers, for more info check out our documentation

Simple example

This script will run the default pipeline:

import cognee
import asyncio


async def main():
    # Add text to cognee
    await cognee.add("Natural language processing (NLP) is an interdisciplinary subfield of computer science and information retrieval.")

    # Generate the knowledge graph
    await cognee.cognify()

    # Query the knowledge graph
    results = await cognee.search("Tell me about NLP")

    # Display the results
    for result in results:
        print(result)


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

Example output:

  Natural Language Processing (NLP) is a cross-disciplinary and interdisciplinary field that involves computer science and information retrieval. It focuses on the interaction between computers and human language, enabling machines to understand and process natural language.

Our paper is out! Read here

Cognee UI

You can also cognify your files and query using cognee UI.

Try cognee UI out locally here.

Understand our architecture

Demos

1. Cogwit Beta demo:

Cogwit Beta

2. Simple GraphRAG demo

Simple GraphRAG demo

3. cognee with Ollama

cognee with local models

Code of Conduct

We are committed to making open source an enjoyable and respectful experience for our community. See CODE_OF_CONDUCT for more information.

💫 Contributors

Sponsors

Thanks to the following companies for sponsoring the ongoing development of cognee.

• GitHub's Secure Open Source Fund

Star History