Skip to content
MCP server

Nautilus Compass Memory

Local-first long-term memory layer for AI agents, served over MCP

Pricing
Open source
Platforms
CLI, API
Runs
Locally or Remotely
GitHub stars
1.3K
Last commit
today
Language
Python

About Nautilus Compass Memory

Nautilus Compass Memory is an open-source MCP server for knowledge and memory. It can run locally on your machine or as a hosted remote server.

Nautilus Compass Memory stores observations and lets AI agents recall them later, with signed verification receipts that can be recomputed. The README lists 17 MCP tools, with core ones for writing observations, hybrid semantic and keyword recall, session search, profile and drift checks.

It installs from PyPI and runs over stdio, with a hosted streamable-HTTP endpoint also in the registry. It works with MCP clients such as Claude, Cline, Cursor, Continue.dev and Zed. The project reports its own benchmark comparisons against another memory product, which are the authors' claims.

The site also mentions paid judging reports, and the license is described as modified MIT.

Install

Run in a terminal

claude mcp add --transport http nautilus-compass https://compass.nautilus.social/mcp/

Add to Settings → Connectors → Add custom connector

https://compass.nautilus.social/mcp/
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "nautilus-compass": {
      "url": "https://compass.nautilus.social/mcp/"
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "nautilus-compass": {
      "type": "http",
      "url": "https://compass.nautilus.social/mcp/"
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "nautilus-compass": {
      "serverUrl": "https://compass.nautilus.social/mcp/"
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "nautilus-compass": {
      "type": "http",
      "url": "https://compass.nautilus.social/mcp/"
    }
  }
}

Tools this server provides (5)

Taken from the project documentation; the running server may differ.

ingest_obs
Write an observation with anchor and drift signal
recall
Hybrid semantic and keyword memory search
session_search
Search session logs by time
profile
Aggregate work profile
drift_check
Check a prompt for drift

What it does

Long-term recall
Hybrid semantic and keyword search over stored memories.
Observation ingestion
Writes observations with an anchor and drift signal.
Drift detection
Flags drift against stored context.
Verification receipts
Signed receipts that can be recomputed.
Local-first
Designed to run on your machine.
Cross-agent use
Shared across several agents and MCP clients.

Good for

  • Give a coding agent persistent memory
  • Search past session logs
  • Detect drift between prompts and prior context
  • Share memory across agents

Pricing

Open source

Open source (modified MIT per site); paid judging reports are separate

Prices change. Check compass.nautilus.social before you buy.

Nautilus Compass Memory FAQ

How many tools does it expose?

17 per the README; seven core tools are listed.

How is it installed?

From PyPI as nautilus-compass, over stdio, or via the hosted endpoint.

Is it free?

The project is open source; some separate judging reports are paid, and the license is unclear in the metadata.

Is Nautilus Compass Memory free?

Yes. Nautilus Compass Memory is open source, so you can use and self-host it for free.

How do I add Nautilus Compass Memory to Claude Code?

Run this command in your terminal: claude mcp add --transport http nautilus-compass https://compass.nautilus.social/mcp/

What are the best alternatives to Nautilus Compass Memory?

The closest alternatives to Nautilus Compass Memory are Codebase Memory, XERJ, Claude Code Ultimate Guide, note1 and MedXpert 图书馆连接器(通用版).

Alternatives to Nautilus Compass Memory

  1. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  2. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  3. MCP server for searching the Claude Code Ultimate Guide from any MCP client

    Runs locally
    • Open source
    • 6.5K GitHub stars
    • AI models and LLM tooling
  4. note1

    MCP server

    MCP server giving AI assistants search and summaries across your note1 meetings

    • 4 tools
    • Remote
    • Free trial
    • 1 GitHub star
    • Productivity
  5. Context Nest

    MCP server

    Governed, versioned context vault that Claude, Cursor and other MCP clients read from

    Not checked yet
    • 4 tools
    • Remote
    • Freemium
    • Knowledge and memory