cognee-community

Install and register community database adapters, connectors, and pipelines for cognee.

30.4k|3.0k|Updated Aug 16, 2023
One-click install
npx skills add https://github.com/topoteretes/cognee --skill cognee-community
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cognee-community
Source: https://github.com/topoteretes/cognee/tree/main/.claude/skills/cognee-community
Command: npx skills add https://github.com/topoteretes/cognee --skill cognee-community

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Cognee core ships with a limited set of built-in database providers and data sources, so teams needing Qdrant, Milvus, Slack, Notion, or other backends must find, install, and correctly register community packages from the separate cognee-community monorepo.

Core Features & Use Cases

  • Database Adapter Registration: Guides installing vector, graph, and hybrid adapters (Qdrant, Milvus, Weaviate, FalkorDB, Memgraph, DuckDB, and more) and importing their register module before cognee initializes any engine.
  • Data-Source Connectors: Explains wiring dlt-based connectors for Slack, Gmail, Notion, Confluence, and Google Drive directly into cognee.remember() with dedicated datasets.
  • Contribution Workflow: Documents how to structure, test, and publish a new package to the cognee-community repo, including the shared contract test suite.
  • Use Case: A team wants cognee to store embeddings in their existing Qdrant instance; this Skill walks through installing cognee-community-vector-adapter-qdrant, importing its register module, and setting the vector DB config.

Quick Start

Ask the assistant to configure cognee to use Qdrant as its vector database using the community adapter package.

Frequently Asked Questions about cognee-community

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I use Qdrant with cognee?

Install cognee-community-vector-adapter-qdrant, then import its register module before cognee touches any engine. Set vector_db_provider to "qdrant" with the URL and key in config.set_vector_db_config.

How do I ingest Slack data into cognee?

Use the cognee-community Slack connector, which exposes a dlt source you pass directly to cognee.remember(). Give the connector its own dedicated dataset so snapshot sync and forget-on-delete work correctly.

Why does cognee raise "Unsupported vector database provider"?

This error occurs when you set a community provider name without importing the package's register module first. The register import calls use_vector_adapter or use_graph_adapter, which makes the provider name valid.

Which community adapters work with cognee access control enabled?

Only adapters shipping a dataset database handler work with ENABLE_BACKEND_ACCESS_CONTROL=true: qdrant, moss, singlestore, turbopuffer, falkordb, arcadedb, and helixdb. All other community adapters require setting ENABLE_BACKEND_ACCESS_CONTROL=false.

How do I contribute a new adapter to cognee-community?

Branch from main, create a package under packages/<family>/<name>/ with pyproject.toml, README, examples, and tests. Implement VectorDBInterface or GraphDBInterface, expose a register.py, and run the shared contract tests in packages/shared/contract_suite.