Your coding standards, always in context

An open-source MCP server that gives your AI assistant semantic search over your personal documentation and coding guidelines. 100% offline.

Get Started View on GitHub

terminal

$ c42 add ./my-standards --name rules --priority 0.9

✓ Source 'rules' added

$ c42 index

✓ Indexed 42 chunks from 1 source

$ c42 serve

✓ MCP server running▌

The problem

Standards get lost

Your AI assistant doesn't know your team's coding conventions. Every session starts from zero.

Docs are scattered

Guidelines live in wikis, READMEs, and Notion pages. None of it reaches your AI tools.

Context windows fill up

Pasting entire docs into prompts wastes tokens and dilutes focus.

Code leaks to the cloud

Most AI tools send your code to external servers. In regulated industries, that's a compliance risk you can't afford.

The solution

Context42 indexes your documentation locally and serves the most relevant chunks to your AI assistant via MCP — automatically, semantically, and privately.

Your Docs

c42 index

Vector DB

MCP

Claude

Features

100% Offline

Runs entirely on your machine. No API calls, no cloud, no data leaving your computer.

Semantic Search

Finds relevant content by meaning, not just keywords. Powered by local embeddings.

Priority System

Weight your personal rules higher than reference docs. Control what matters most.

MCP Native

Built for the Model Context Protocol. Works with Claude Desktop out of the box.

Multi-format

Index Markdown and reStructuredText files. More formats coming soon.

Smart Indexing

Content-hash based deduplication. Re-run safely without duplicating data.

Security & Privacy

Your code never leaves your machine

Context42 processes everything locally. No cloud, no APIs, no third-party access — so your source code and documentation stay exactly where they should.

Zero data transmission

No outbound network calls. Your code, docs, and embeddings are processed and stored entirely on your device.

Local-only embeddings

The AI model runs on your CPU. No tokens sent to external APIs — the embedding model is downloaded once and runs offline forever.

No telemetry, no tracking

Context42 collects nothing. No usage analytics, no crash reports, no phone-home behavior. Fully transparent, open-source codebase.

Works air-gapped

After initial setup, no internet connection is needed. Perfect for restricted networks and secure environments.

Your machine

Source code & docs

Embedding model

Vector index

No data out

Internet

Helps your team meet compliance requirements

GDPR No cross-border data transfer

HIPAA No third-party data handling

SOC 2 Full data access control

ISO 27001 Data stays in your perimeter

ITAR On-premises processing

CMMC CUI under your control

Context42 helps support compliance by keeping all data local. Compliance certification depends on your organization's overall security posture.

How it works

Add your sources

Point Context42 to your documentation directories.

c42 add ./my-standards --name standards --priority 0.9
c42 add ./docs --name reference --priority 0.5

Index

Context42 chunks your docs and creates local vector embeddings.

c42 index

Manage your sources

List, adjust priorities, search, or remove sources at any time.

c42 list
c42 set-priority standards 1.0
c42 search "error handling"
c42 status

Connect to Claude

Start the MCP server and configure Claude Desktop.

c42 serve

Get started in under a minute

# Install with pipx (recommended), uvx, or pip
pipx install context42-io

# Add your docs, index, and serve
c42 add ./my-docs --name my-docs
c42 index
c42 serve

Claude Desktop configuration

claude_desktop_config.json

{
  "mcpServers": {
    "context42": {
      "command": "c42",
      "args": ["serve"]
    }
  }
}

Open Development

Roadmap

Context42 is built in the open. See what's planned, suggest features, and help shape the project.

Up next

Git clone sources

Add any Git repository as a source — public or private. Context42 clones it locally and indexes the content, no manual download needed.

Planned

File watcher

Automatically re-index sources when files change on disk. A local watchdog monitors your directories so your embeddings are always up to date.