Skip to content

Quickstart

Prerequisites

  • Node.js 18+ or Bun — Moneypenny runs on Bun but installs via npm/npx
  • An LLM API key — Anthropic (default), OpenAI, or Google

1. Install and Initialize

Terminal window
npx moneypenny init

This creates:

  • .mp/ — project directory with agent configuration
  • .mp/mp.db — the SQLite database (sessions, code index, policies, audit)
  • .mp/workspace.sqlite — hybrid search index over your codebase
  • A default agent ready to use

2. Configure Your API Key

Terminal window
mp config set anthropic_api_key sk-ant-...

Or use environment variables:

Terminal window
export ANTHROPIC_API_KEY=sk-ant-...

3. Verify Setup

Terminal window
mp doctor

Doctor checks: API key configured, SQLite extensions loaded, embeddings working, index status, MCP server ready.

4. Start Coding

Interactive chat:

Terminal window
mp chat

One-shot message:

Terminal window
mp chat "refactor the auth module to use JWT instead of sessions"

The agent reads your code, proposes changes, writes files, and runs tests. Everything is persisted in .mp/mp.db.

5. Search Your Code

Terminal window
mp search "where do we handle rate limiting?"

Hybrid search combines BM25 keyword matching with vector semantic search, fused via Reciprocal Rank Fusion. The codebase is indexed incrementally — first index takes 30-60 seconds, subsequent updates are sub-second.

6. Add Moneypenny to Your IDE (Optional)

Terminal window
mp setup cursor

Writes .cursor/mcp.json. Restart Cursor to activate. Moneypenny’s code search, policy engine, and audit log are now available as tools.