Getting Started

Get up and running with Chub in under 5 minutes. This guide covers the three pillars — context, tracking, and team knowledge — and how to set them all up.

Install

Pick your preferred method:

npm

npm install -g @nrl-ai/chub

pip

pip install chub

Cargo

cargo install chub

Homebrew

brew install nrl-ai/tap/chub

See the full Installation guide for binary downloads and platform-specific instructions.

Verify it works:

chub --version

Search for docs

Chub serves curated API documentation from a public registry of 1,553+ docs. Search for anything:

chub search "stripe payments"

  1. stripe/api            Stripe API reference           ★ 0.92
  2. stripe/webhooks       Stripe webhook handling         ★ 0.78
  3. stripe/checkout       Stripe Checkout integration     ★ 0.71

Fetch a doc

Grab a specific doc by ID. Use --lang to get language-specific content:

chub get openai/chat --lang python

This outputs the full markdown doc — ready to be consumed by an AI agent or read by a human.

# Other examples
chub get stripe/api --lang javascript
chub get nextjs/app-router --version 15.0
chub get openai/chat --lang python --version 4.0

List all docs

chub list

Use --json with any command for machine-readable output:

chub list --json

Set up MCP for your AI agent

Chub includes a built-in MCP (Model Context Protocol) server. This is how AI agents like Claude and Cursor access docs automatically.

Claude Code (.mcp.json)

{
  "mcpServers": {
    "chub": {
      "command": "chub",
      "args": ["mcp"]
    }
  }
}

Cursor (.cursor/mcp.json)

{
  "mcpServers": {
    "chub": {
      "command": "chub",
      "args": ["mcp"]
    }
  }
}

Once configured, your AI agent can search and fetch docs without any manual commands.

Initialize a project

Set up team sharing by creating a .chub/ directory in your project:

chub init

This creates:

my-project/
├── .chub/
│   ├── config.yaml        # Project config
│   ├── pins.yaml          # Pinned docs
│   ├── annotations/       # Team-shared annotations
│   ├── context/           # Custom project docs
│   └── profiles/          # Named context profiles

Auto-detect dependencies

Use --from-deps to scan package.json, Cargo.toml, requirements.txt, etc. and auto-pin matching docs:

chub init --from-deps

Commit .chub/ to git so the whole team shares the same context. Personal settings stay in ~/.chub/.

Pin docs for your team

Lock specific doc versions so every team member and AI agent uses the same reference:

chub pin add openai/chat --lang python --version 4.0 --reason "Use v4 streaming API"
chub pin add stripe/api --lang javascript

List and fetch pinned docs:

chub pin list          # list all pins
chub pin get           # fetch all pinned docs at once

Track your AI usage

Enable session tracking to see tokens, costs, and tool usage across any AI agent:

chub track enable                  # auto-detects Claude Code, Cursor, Copilot, etc.

Then use your AI agent as normal. After a session:

chub track status                  # see active session
chub track report                  # aggregate usage report (costs, tokens, models)
chub track dashboard               # web dashboard at localhost:4243

Session summaries are shared with your team via git. Full transcripts stay local. See AI Usage Tracking for details.

Config inheritance

Chub uses a layered config file system — no layer is required. Settings cascade from personal defaults through project config to active role profiles:

~/.chub/config.yaml          # Personal defaults (machine-wide)
    ↓ overridden by
.chub/config.yaml            # Project config (git-tracked, shared with team)
    ↓ overridden by
.chub/profiles/<name>.yaml   # Role/task profile (e.g. backend, frontend)

What to learn next

Now that you have Chub installed, explore the features that matter to your workflow:

If you want to...	Read
Understand the vision	Why Chub
Context
Lock doc versions for your team	Doc Pinning
Give different roles different context	Context Profiles
Add custom project docs	Project Context
Auto-detect deps and pin docs	Dep Auto-Detection
Self-Learning
Share team knowledge in git	Annotations & Self-Learning
Rate doc quality for maintainers	Feedback
Sync CLAUDE.md / .cursorrules	Agent Config Sync
Tracking & Analytics
Track sessions, tokens, and costs	AI Usage Tracking
Monitor doc freshness	Snapshots & Freshness
Reference
See all CLI commands	CLI Reference
Configure Chub	Configuration
Connect AI agents via MCP	MCP Server
Author docs and skills	Content Guide
Build a custom private registry	Self-Hosting
Compare with Context Hub or Context7	Comparisons