What problem does it solve?
Running cognee locally requires configuring databases, LLM providers, and multiple services; this Skill guides you through starting cognee with Docker, from a single prebuilt container to the full compose stack with UI, MCP, Postgres, and Neo4j.
Core Features & Use Cases
- Prebuilt Image Quickstart: Run the cognee/cognee:main image with only an LLM_API_KEY, then verify with the /health endpoint and interactive docs at port 8000.
- Memory API Usage: Ingest and query data through /api/v1/remember and /api/v1/recall, including search_type routing behavior and snake_case/camelCase field aliases.
- Full Stack Profiles: Bring up optional compose profiles for the frontend (port 3000), MCP server (port 8001), Postgres/pgvector, and Neo4j.
- Use Case: You want to try cognee without cloning the repo: save a minimal docker-compose.yml, export your OpenAI key, run docker compose up, and immediately remember and recall a note via curl.
Quick Start
Ask the AI to start cognee locally with the prebuilt Docker image and show the curl commands to remember a note and recall it.