Skip to main content

1. Get your API key

Go to Settings in the app. Under API Keys, click Generate. You’ll get a pb_... key and a ready-to-use snippet.

2. Copy the quickstart snippet

The modal gives you everything your agent needs in one block:

3. Give it to your agent

Paste the snippet into your Claude Code session, Cursor config, or any LLM agent. The agent reads the LLM Guide first, then uses the API to understand your project. Claude Code example:

4. Agent reads the tree

The agent calls GET /api/v1/nodes?projectId=... to see your full planning tree, and GET /api/v1/iterations?projectId=...&current=true to find the active iteration. It knows what to build and what proof to produce.

5. Agent writes back

As the agent works, it updates the tree:

6. You review

Open the Delivery Map in the app. See which jobs the agent has completed, what’s still open, and how your bets are progressing across iterations.

What’s next

API Reference

Full endpoint docs, nodes, search, mutations, iterations.

LLM Operations Guide

The canonical reference your agent reads. Data model, field rules, all endpoints.