README.md

Get started

This is a community-maintained plugins repo, where you can find various implementations of adapters, custom pipelines, etc. You can check out our core repo on how to get started with cognee.

You can install the chosen community package in two different ways:

Install using pip

Install the chosen community packaging using uv pip install ...

uv pip install cognee-community-vector-adapter-qdrant

Install using UV or poetry locally with all optional dependencies

Navigate to the packages folder and the adapter of your choice, and run either of the following commands:

uv sync --all-extras
# OR
poetry install

OR

poetry install

You will need an LLM API key (OpenAI by default) in order to run cognee with these adapters. Before importing cognee, make sure to define your key like so:

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

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

Run an example to verify installation

Navigate to the package directory of your choice and run the example, usually found in the examples directory. You can run them either via uv, or poetry:

uv run python ./examples/example.py
# OR
poetry run python ./examples/example.py

Supported Database Adapters

Package Name	Type	Description
cognee-community-vector-adapter-azure	Vector	Azure AI search vector database adapter for cognee
cognee-community-vector-adapter-milvus	Vector	Milvus vector database adapter for cognee
cognee-community-vector-adapter-opensearch	Vector	Opensearch vector database adapter for cognee
cognee-community-vector-adapter-pinecone	Vector	Pinecone vector database adapter for cognee
cognee-community-vector-adapter-qdrant	Vector	Qdrant vector database adapter for cognee
cognee-community-vector-adapter-redis	Vector	Redis vector database adapter for cognee
cognee-community-vector-adapter-valkey	Vector	Valkey vector database adapter for cognee
cognee-community-vector-adapter-weaviate	Vector	Weaviate vector database adapter for cognee
cognee-community-graph-adapter-memgraph	Graph	Memgraph graph database adapter for cognee
cognee-community-graph-adapter-networkx	Graph	Networkx graph database adapter for cognee
cognee-community-graph-adapter-spanner	Graph	Google Cloud Spanner (Spanner Graph) adapter for cognee
cognee-community-graph-adapter-turingdb	Graph	TuringDB graph database adapter for cognee
cognee-community-graph-adapter-pggraph	Graph	pgGraph adapter for cognee (Postgres + graph extension, experimental)
cognee-community-hybrid-adapter-duckdb	Hybrid	DuckDB hybrid database adapter for cognee
cognee-community-hybrid-adapter-falkor	Hybrid	FalkorDB hybrid database adapter for cognee

Custom Packages

Custom packages are also a part of this repo, containing, for example, custom pipelines, tasks, and retrievers. Every pipeline has its own package, as well as every retriever. Tasks are grouped so that all mutually relevant tasks are contained in one package (i.e. all tasks used in one pipeline are packaged together).

Current Custom Packages

Package Name	Type	Description
cognee-community-pipeline-codify	Pipeline	Custom codify pipeline package
cognee-community-retriever-code	Retriever	Custom CODE retriever package
cognee-community-tasks-codify	Task	Custom codify tasks package
cognee-community-tasks-scrapegraph	Task	Web scraping tasks powered by ScrapeGraphAI
cognee-community-tasks-exa	Task	Web search tasks powered by Exa

Repository Structure

• All packages are located in the packages directory.
• Each package must include:
  • A README.md file with installation and usage instructions.
  • An examples directory containing an example.py file demonstrating how to use the plugin with Cognee.
  • A tests directory containing tests at least a bit more detailed than the example.
• When adding new adapters or custom packages, follow the structure of the existing packages.

Contributing

We welcome contributions from the community! Your input helps make Cognee better for everyone. See CONTRIBUTING.md to get started.