Skip to main content
Ref Plans is a tool for writing, iterating, and collaborating on plans for coding agents. It’s a persistent, document-first approach to agent orchestration. Instead of ephemeral chat sessions that disappear when you close a tab, Plans give you a durable workspace where context survives across sessions, teammates can collaborate, and multiple agents can work in parallel.

​
How it works

1

Write an outline

Start with your notes, a Linear ticket, a GitHub issue, or just a rough idea. Drop it into a new plan at plan.ref.tools.
2

Research and refine with AI

Ref’s agent researches your codebase, reads relevant files, and breaks your outline into structured tasks with implementation steps and verification criteria.
3

Send tasks to coding agents

When you’re happy with the plan, send individual tasks to Claude Code, Cursor Background Agents, Codex, Devin, or any other agent. They report progress back to Ref as they work.

​
Why use Ref Plans?

  • Persistent docs, not ephemeral chat — Plans live as documents you can revisit, share, and iterate on over days or weeks. Context doesn’t vanish when you close a tab.
  • Human review before code — Catch architecture issues and align on approach before agents write a single line. Review plans like you’d review a design doc.
  • Parallel agent orchestration — Send different tasks to different agents at the same time. Track their progress from one place.
  • Context that survives sessions — Plans maintain the full history of research, decisions, and progress. New agents pick up where the last one left off.
  • Web-based and mobile-friendly — Start a plan on your laptop, review it on your phone, launch agents from anywhere.

Quick Start

Create your first plan in minutes.

Best Practices

How to write plans that get great results.