AI software factory for Claude Code and Codex
Put your coding agent to work across your whole backlog.
Use the Claude Code or Codex subscription you already pay for. ampm runs the engineering process around it: planning the change, building in an isolated sandbox, testing and reviewing the result, and opening the pull request. You keep the decisions and every receipt.
A kanban board with four lanes — Ideating, Planning, Building and Shipping — carrying tasks from a fresh idea through to a pull request waiting for review.
What actually happens between midnight and your first coffee
- 11:47 PMYou
Two sentences and a light switch.
“The CSV export falls over on big accounts — it buffers the whole thing in memory.” You file it the way you'd tell a colleague, in whatever words you have at midnight. No template, no ticket taxonomy, no estimating poker.
- 3:12 AMThe factory
Nobody is awake. Everybody is working.
A refiner turns your sentence into acceptance criteria. A planner writes the plan. A builder implements it in an isolated sandbox, the tests run, a reviewer pushes back, the builder goes again. Three rounds. The reviewer stops finding things.
- 8:30 AMYou
One PR. One question. Coffee.
A pull request, already reviewed and tested, waiting for your yes. Beside it, one honest question: the fix changes a response's shape, and the reviewer wants you to pick. You answer it in a sentence. The rest of the night needed nothing from you at all.
Have a poke around
See the whole factory at work
There is no dashboard to configure and no workflow builder to learn. You file work, you answer what's blocked on you, and you watch the rest happen. Here it is, with invented data — press + New task in the inbox, approve the breakdown, and watch three ordered tasks appear and start building.
Scroll a little further to load the demo
The board
Every piece of work, at the exact stage it's in. Positions aren't decoration — a card's column IS its state.
The inbox
One queue of everything blocked on you, each ask carrying the plan, the diff and the transcript behind it.
Activity
Who's working right now, on what, and what each run cost — live, with every finished run still on the record.
The inbox
Nothing merges unless you want it to.
The decisions that matter — the plan before anyone writes code, the call only you can make, the merge into your main branch — wait for your yes. Not because the agents are asked politely to stop: the board simply has no way past them without an answer. Hand off the ones you trust it with; keep the rest.
Every ask arrives in one queue with its receipts attached: the plan, the diff, the PR, the full transcript of what the agent actually did. Approve it, send it back with a reason the crew has to answer, or just reply to the question. Then close the laptop again.
- planplan · diff · transcript
Is this the right approach before anyone writes code?
- questionplan · diff · transcript
Keep the old endpoint alive for a release, or cut over now?
- mergeplan · diff · transcript
Does this change go into your main branch?
The lifecycle
A process, not another chat window
A chat forgets everything the moment you close the tab, which is why it can't be trusted with anything longer than an afternoon. ampm keeps a real database instead: every task is a row with its plan attached, its history intact, and its place in the tree of work it came from. Retries don't overwrite anything — they land beside the last attempt, because you should be able to ask what happened last time.
Plan before you build, test and review before you merge — that discipline wasn't invented here. It's how the best San Francisco engineering orgs actually ship, distilled from fifteen years as an engineer inside them and codified into machinery. You get the institutional knowledge it takes a career to absorb, working for you from day one — and holding the bar even on the nights you wouldn't have bothered.
- Planningyour call
- Implementing
- Testing
- Reviewing code
- Mergingyour call
Safe to leave running on your real repos
Your main branch is never the test bench
Agents work in isolated sandboxes on their own worktrees, with no path to your credentials and no way to push anywhere but a branch. The output of a night's work is always a pull request.
Read exactly what it did
Every turn keeps its full transcript. When a change surprises you, you don't reconstruct the reasoning from the diff — you open the run and read it.
Nothing quietly disappears
A retry is a new run beside the old one, not a rewrite of it. Failures park with the reason attached and wait for you, instead of looping until something gives.
One factory, all your repos
A workspace spans the repos agents change and the ones they only read for context — a design system, shared docs, an API they call. A change that needs a migration in one and a client update in another is one piece of work with two children, not two tickets you keep in sync yourself.
Questions
Which coding agents does ampm work with?
Claude Code and Codex. ampm runs them on the subscription you already pay for — no API keys and no second bill for tokens — and when one account hits its usage limit, work moves to a backup account without losing anything.
Does ampm merge code on its own?
Nothing merges unless you want it to. Plans, questions and merges wait in your inbox with the plan, diff, pull request and full transcript attached, and you decide which decisions to keep and which to hand off.
Where does the code run, and is my repo safe?
Every task runs in an isolated sandbox on its own worktree, on workers you choose — your laptop, a server you own, or a cloud machine — and can only push to a branch. The output is always a pull request into your repo.
How is this different from chatting with a coding agent?
A chat forgets everything when you close the tab and needs you watching it. ampm keeps every task in a database with its plan and history, runs a real process — plan, implement, test, review, merge — and only asks you when a decision is actually yours.
Can one workspace cover several repos?
Yes. Add the repos agents should change, and add more as read-only context — a design system, an API they call, product docs — so their work matches what's around it. One change can span several repos as one piece of work.
How do I get access?
Join the wait list with Google or GitHub. We're letting people in a few at a time, in the order they joined, and we'll email you when it's your turn.
Early access
Get your factory running
We're letting people in a few at a time. Join the wait list with Google or GitHub and we'll email you when it's your turn. Run it on our cloud or on infrastructure you control — your repos, your subscriptions, workers on machines you choose.