rev0 is a kanban board on your machine that runs Claude Code, Codex and Gemini, one git worktree per ticket, and stops at you for every decision.

You are the bottleneck now. Agents run tens at a time; you still read one diff at a time.

Early access · v0.3.0Join the waitlist

Sign in, get on the list, and we email you when your access is ready. Free beta for personal and professional use, including commercial work: licence.

rev0 is in beta: free on your own machine, accounts for sync and phone open in waves from the waitlist.

AK-1503 · build me a snake game window
The rev0 board: two agents at work on their own branches, a plan waiting for sign-off, and the whale plugin swimming across.
rev0 planned, captured and cut this film. Motion graphics explain the ideas; the screens are captures of its current board.

01The ticket

A ticket crosses the board. You are in the middle of it.

Every lane is a place you can stop it, read what it did, and change your mind.

  • Todo
  • In Progress
  • Your turn
  • In Review
  • Done
The rev0 board: tickets in Backlog and To Do, two cards waiting in Your turn that say why they wait, one in review and two done.
A capture of the board. Each card in Your turn says why it waits.

Its own worktree and branch

Five agents at once never touch the same files, and one reviewed merge reaches your base branch.

The engine is a setting

  • Claude Code
  • Codex
  • Gemini
  • Cursor

Your subscription, no API bill

No API key and no meter: five agents cost what your plan already costs.

02The plan

The plan is a page you can click.

Click any element, leave a note, and it goes back to the agent pinned to that element. The one below is live.

SPA plan · awaiting sign-off
Goal Move sessions out of the app process and into Redis
Option A Redis, with a dual-write window. Two days, no downtime.
Option B Postgres table plus an index. Half a day, slower reads.
Rollout
Risk Sessions issued before the cut over will not carry the new claim

    Click any part of the plan above.

    Three shapes of plan

    Markdown when that is enough, a graph of steps you can rearrange, or a full interactive page.

    Reject one part, keep the rest

    You argue with one step, the agent answers that step, and the rest of the plan stands.

    An interactive plan on the board, with a note being written on one of its cards: keep this card above the fold on phones.
    A capture of a plan on the board, a note being pinned to one card.

    03In parallel

    A pipeline you can see before it runs.

    A Loop is a graph you edit and approve, and it runs in exactly that order. Give it a schedule and the rounds you make by hand are already done.

    Nightly, 02:00 Test coverage Performance Security review Merge findings Human approval Post the digest
    Three nodes with no path between them run at the same time.

    Loops · on a schedule

    • The GitHub pull requests waiting on you, prerevieweddaily 09:00
    • Your Jira issues and Slack threads, digested for standupMon to Thu 12:45
    • Outdated and vulnerable packages on npm and PyPIMondays 09:00
    • The work you already did on the board, with what to automate nextFridays 17:00

    Each run is a ticket of its own, with its report attached.

    Gates in the graph

    A human approval before anything irreversible, or a reviewer from a different model family.

    Ready when you need it

    Set the time you need the result and a head start. The run begins early.

    It reads, you decide

    The ready made reports are read only. Anything that posts or sends waits at a gate.

    04The gate

    It does not run on trust.

    Anything outside the allow-list stops the run and asks you, reasoning attached. What reaches you has already passed your numbers.

    Your turn · needs permission

    I'm requesting set_plugin_source(state="loaded") because that is the last step the ticket asks for. Expected effect: a new Snake tile becomes clickable in App Space. To undo: call it again with state = "unloaded".

    Go on, click one. Nothing happens here. On a real board this is the moment the run stops and waits for you.

    Quality gate · you never see the first draft

    Coverage went from 61 percent to 86 percent.

    • 1wrote parse_manifest()coverage 61%
    • 2gate: 80% is the floor, and 2 lint errorsrejected
    • 3added two tests, fixed the lintcoverage 86%

    It corrected itself. You get the version that passes.

    Deny always wins

    Allow, deny and ask lists per board. A ban is absolute and an ask always reaches a human.

    The agent argues its own case

    Every request arrives with what it wants, why, and how to undo it.

    Your numbers, only on what it touched

    File length, complexity, coverage and lint. Debt you did not write is left alone.

    05Beyond the repo

    It has hands.

    A browser, the apps on your Mac, your phone, and whatever you teach it.

    browser

    A browser you can watch and take over.

    • Pointed this site's six domains at Vercel in Porkbun's DNS editor
    • Gathered 31 MacBook Pro listings on Leboncoin, eBay and five more sites

    Both are tickets that ran on the author's own board. You sign in once, in its browser: it does not ask for your password in the chat. Any Chrome extension you install loads in every ticket's browser.

    Mac apps

    And the apps on your Mac.

    • Keynote
    • Numbers

    It opens an app, reads its window and uses its menus. The first time it needs an app, it asks you.

    anywhere

    It runs on your phone. Today.

    The same board, installed like an app. Your phone buzzes when an agent stops to ask.

    extend

    Teach it new tricks.

    A plugin adds a command, a panel or a tool. Code that runs on the board asks first, and a plugin is cloned into a directory the board owns, never into yours.

    being built now

    rev0 is trying things nobody else is.

    • Agents that teach, not just deliver
    • Loops that are fully deterministic
    • A hosted version

    An experiment in what supervising agents should feel like.

    Questions

    Before you join the waitlist.

    Is rev0 free?

    Free on your own machine during the beta; accounts for sync and phone open in waves from the waitlist.

    Which coding agents does it run?

    Claude Code by default, Codex, Cursor, Gemini through Antigravity.

    Do I need an API key?

    No new one: rev0 drives the agent CLI you already signed in, on the subscription or key you already use.

    Where does my code run?

    On your machine, in a git worktree per ticket; nothing an agent writes reaches your branch until you merge it.

    Is rev0 open source?

    No. rev0 is in closed beta under the Beta Evaluation License.

    rev0 by cpc.studio

    Every ticket starts at rev 0.

    Provenance

    Built by itself.

    Every line of this page and the tool behind it was written by agents on this board, merged through the gate you just clicked. It runs on your machine, on your own repos.

    Early access · v0.3.0Join the waitlist Read the tutorial

    Sign in, get on the list, and we email you when your access is ready.