Skip to content

Run the cognee Server

Start, check, and connect to a local cognee API server and UI, and pick the right auth mode.

Stars
33,002
Licence
Apache-2.0
Files
1 file

What it does

This skill teaches an AI agent how to run cognee, an open-source memory engine for AI apps, as a local web server with an optional browser UI. It explains how to start the server, check that it is healthy, point the cognee SDK or CLI at it, and choose between multi-user access control and a single-user mode with no login. It also covers common errors such as port conflicts and permission-filtered empty search results.

Great for

  • Developers who want a personal cognee server running on their own machine
  • Teams testing how the SDK or CLI talks to a running cognee API
  • People who need to decide whether authentication should be on or off for a local setup

Skip it if

  • Exposing a server to a network or the internet with auth turned off, since the skill itself says single-user mode is never appropriate for anything exposed
  • Users who want a Docker-only deployment, which the file redirects to a separate cognee-docker skill

Try asking

  • “Start the cognee server and UI on my machine and check that it is healthy”
  • “Connect my Python script to my local cognee server instead of embedded mode”
  • “Why do all my cognee API calls return 401 and how do I fix it for local dev?”
  • “Show me how to run the graph visualization without the full UI”
Step-by-step guide: how to use Run the cognee ServerInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.

    cognee-server.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip cognee-server.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip cognee-server.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Keep ENABLE_BACKEND_ACCESS_CONTROL at its default true unless this is a private single-user dev machine.
  • Use the interactive docs at localhost:8000/docs or GET /openapi.json to confirm which routes your version exposes.
  • If search returns an empty list instead of an error, check the calling user's dataset permissions before assuming no data exists.

What's inside

  • SKILL.md

Redistributed unmodified from topoteretes/cognee under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.