MidOS - MCP Community Library

io.github.MidOSresearchPublished by GitHub account MidOSresearchv2026.2.27Updated Oct 24, 2026

Curated knowledge for AI agents — 19K chunks, 121 skills, semantic search via MCP

VerifiedSTDIODesktop onlyKnowledge & MemoryAI & ML

Overview

AI-generated overview

MidOS exposes a local MidOS knowledge base, skills and research library to an AI assistant through MCP tools for search and retrieval.

What you can do
Search curated knowledge chunks across domains, with full content returned rather than truncated excerpts when available through the vector store option. Look up reusable skill patterns across technology stacks. Retrieve validated improvements, verified patches and corrections, and state-of-the-art reference entries. Run semantic search over the knowledge base using LanceDB and embeddings. Check server health and knowledge base statistics, and generate client configuration files.
When to use it
Useful when you already have a MidOS installation and want an assistant to draw on its curated knowledge, skills and research entries while working in a supported client such as Claude Code, Cursor or Cline. Less relevant if you have no local MidOS knowledge base to point it at.
Requirements
Local Python 3.10+ process installed from PyPI, with FastMCP 2.x. A local MidOS knowledge base is required, located via the MIDOS_ROOT environment variable or auto-detection. The optional vector extra adds LanceDB and embeddings. An optional MIDOS_API_KEY enables authenticated access; the free Dev tier needs no key.
Before you install
The Pro tier costs $10 per month and unlocks write operations and additional tool packs. Invalid or expired API keys silently fall back to the free Dev tier, so check which tier is active. The free tier is limited to 500 queries per month. The server reads a local knowledge base directory, so point MIDOS_ROOT only at a directory you intend to expose.

Installation

In SourceWeft

  1. Open MidOS - MCP Community Library in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

midos-mcp

MidOS MCP Community Library — Knowledge, Skills, and Research as MCP tools.

Install once, connect to any AI client.

Install

bash
pip install midos-mcp

Requirement: MidOS knowledge base must be available locally. Set the MIDOS_ROOT environment variable to point to your MidOS installation directory (the one containing knowledge/ and modules/).

bash
export MIDOS_ROOT=/path/to/midos

Quick Start

bash
# Start server (stdio — default for Claude Code, Cursor, etc.)midos-mcp serve
# Start with HTTP transportmidos-mcp serve --http
# Generate config for your AI clientmidos-mcp config --generate claudemidos-mcp config --generate cursor
# Check healthmidos-mcp health

Commands

midos-mcp serve

Start the MCP server with your preferred transport.

bash
midos-mcp serve              # stdio (default)midos-mcp serve --http       # Streamable HTTP on 127.0.0.1:8419midos-mcp serve --sse        # Legacy SSE transportmidos-mcp serve --http --host 0.0.0.0 --port 9000
FlagTransportUse Case
--stdiostdioClaude Code, Cursor, Cline (default)
--httpStreamable HTTPWeb clients, remote access
--sseServer-Sent EventsLegacy clients

midos-mcp config

Generate ready-to-paste JSON config for any supported MCP client.

bash
midos-mcp config --generate claude    # Claude Desktop / Claude Codemidos-mcp config --generate cursor    # Cursor IDEmidos-mcp config --generate cline     # Cline (VS Code)midos-mcp config --generate windsurf  # Windsurf IDEmidos-mcp config --generate continue  # Continue.devmidos-mcp config --generate zed       # Zed Editormidos-mcp config --generate opencode  # OpenCodemidos-mcp config --generate http      # Generic HTTP client
# Write to filemidos-mcp config --generate claude -o mcp-config.json

midos-mcp health

Check server health: knowledge base stats, vector store status, dependencies.

bash
midos-mcp health          # Human-readable outputmidos-mcp health --json   # JSON output for scripts

midos-mcp keys

Manage API keys for tiered access control.

bash
midos-mcp keys generate --name "my-app" --tier devmidos-mcp keys listmidos-mcp keys revoke midos_sk_abc

Tiers: dev (free), pro ($10/mo), admin.

Configuration

Environment Variables

VariableDefaultDescription
MIDOS_ROOTauto-detectPath to MidOS installation (required)
MIDOS_HOST127.0.0.1HTTP server bind host
MIDOS_PORT8419HTTP server bind port
MIDOS_TRANSPORTstdioDefault transport

Root Detection

The server locates the MidOS knowledge base:

  1. MIDOS_ROOT environment variable (recommended)
  2. Auto-detect: walks up from package location looking for CLAUDE.md + knowledge/
  3. Fallback: current working directory

If you installed via pip, auto-detect won't find the knowledge base. Either:

bash
# Option A: Clone the repo and point to itgit clone https://github.com/MidOSresearch/midos.gitexport MIDOS_ROOT=./midos
# Option B: Point to an existing MidOS directoryexport MIDOS_ROOT=/path/to/your/midos

Verify with midos-mcp health — it shows the detected root and knowledge counts.

Supported Clients

ClientConfig FileTransport
Claude Desktopclaude_desktop_config.jsonstdio
Claude Code.mcp.jsonstdio
Cursor.cursor/mcp.jsonstdio
Clinecline_mcp_settings.jsonstdio
Windsurf~/.codeium/windsurf/mcp_config.jsonstdio
Continue.dev~/.continue/config.jsonstdio
Zed~/.config/zed/settings.jsonstdio
OpenCodeopencode.jsonSSE

Optional Dependencies

bash
# Vector store support (LanceDB + embeddings)pip install midos-mcp[vector]
# Development toolspip install midos-mcp[dev]

What's Inside

MidOS exposes 68 MCP tools across 2 tiers:

TierToolsAccess
Dev34 toolsFree — no API key needed, full content, 500 queries/month
Pro68 tools$10/mo — security ops, AOTC, orchestration, maker write ops

Capabilities

  • Knowledge — Search 46K+ curated chunks across all domains (full content, no truncation)
  • Skills — 119 reusable patterns across 16+ technology stacks
  • EUREKA — 383 validated improvements with measured ROI
  • Truth — 50 verified patches and corrections
  • SOTA — 140 state-of-the-art convergence ceilings
  • Vector Search — Semantic search via LanceDB + Gemini embeddings (3072-d)

Rate Limits

TierQueries/monthContent
Dev500Full content (no truncation)
Pro100,000Full content + AOTC + ops packs

Invalid or expired API keys silently fall back to Dev tier.

Requirements

  • Python 3.10+
  • FastMCP 2.x
  • MidOS knowledge base (local)

License

MIT

Source: README.md at commit 4642a7d

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v2026.2.27LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts