AgentCron

Your local agent workspace.

Less juggling. More building.

Schedule your agents. Review their work. Keep the decisions yours.

See release status See the use cases Not on sale yet

Free for one project. Pro is $89, once — no subscription and no server to fund.

The AgentCron window: a schedule of five agents — two finished, one waiting on you — a review panel holding an uncommitted diff with its explanation card, pills for the three CLIs with a cost line under them, and a live terminal session along the bottom.
Illustration Your morning control room: scheduled runs, an uncommitted diff, and a question waiting for your decision.
  • You bring the CLIUse the agent subscription you already have. No AgentCron API key, account or token meter.
  • Human review is the destinationRuns return work and records for you to inspect. Where a workflow uses AgentCron's approval command, denial, timeout or an unreachable app all fail closed.
  • Runs locally, leaves receiptsYour schedules and app records stay on your Mac. Connected CLIs and services keep their own network and data policies.

Drives the agent CLIs you already pay for

  • Codex CLIcodex
  • Claude Codeclaude
  • GitHub Copilot CLIcopilot
  • Antigravityagy
  • Plain shellshell

You bring the CLI and its subscription. AgentCron schedules the run, keeps its output, and shows the command evidence that integration can capture.

What kind of work keeps returning?

Choose the closest role. Three audience packs are bundled; the other seven starting sets are assembled from the included picker templates. Every page says which is which.

Indie hacker / founder

The chores ran while you slept. Your morning starts at the decisions.

You are engineering, support, finance and marketing — every hat, one head. The chores are not hard, they are just endless, and they are the reason you shipped nothing this week. Overnight they get read, sorted and drafted, so your day begins at the decision instead of at the discovery.

See release status See pricing

Not released yet. It will need macOS 12.0 or later on Apple Silicon — and the download page says exactly what is still missing.

The morning view: a Needs-you strip with four drafts and one decision, each produced overnight and each one click away.
Illustration 08:00. Four drafts and one decision, all produced while you were asleep. Every one of them waiting on you.

The caveat, up front. Every non-code deliverable arrives as a draft you approve and send. For a company of one you are the only reviewer there is, and that is exactly who this is built for.

Day one: three agents, already written

  • Overnight backlog

    Every night at 2am

    By morning there is a finished, uncommitted diff in a worktree with an explanation beside every hunk.

  • Inbox triage

    Every 30 minutes, business hours

    The inbox read, sorted and the obvious replies drafted — and you pulled in only for what truly needs you.

  • Renewal reminder

    Every morning at 8

    No renewal, expiry or deadline arrives as a surprise, each one filed with its source attached.

One agent comes from the developer pack; the other two are ready-made prompts in the agent picker. All three are switched off until you turn them on. Every agent arrives switched off, with a “Run once” button — nothing fires on its own.

What has to be true first

  • A mail account you are willing to connect over IMAP (read-only)
  • Somewhere to put the non-code project: a scratch folder, not a repo

And the limits, in the same breath

  • macOS only. You bring, install and authenticate one of the supported agent CLIs shown above.
  • Schedules fire while your Mac is awake — AgentCron can hold it awake through a window you set (lid open) — and there is no daemon, no wake-from-sleep and no catch-up for a run that was missed.
  • AgentCron grants none of its own commit, push, send or pay paths by default. The launched CLI remains a real networked process, so separately configured CLI tools and project credentials keep their own risk.

The week that disappeared

Nothing went wrong. A customer needed a reply. A certificate was about to expire. A renewal came up that you would have negotiated if you had noticed. Three support emails needed one paragraph each. A contract needed reading. None of it was hard, none of it took long, and together they took the week.

The expensive part was never the typing. It was the reading — working out which of the forty things in front of you were the four that mattered, and reconstructing enough context to answer each one.

That is the part that happens overnight. You wake up to four drafts and one decision, read them, and send. Five minutes of approving instead of two hours of finding out.

Your agent drafts. You send. That is the whole design.

Agents here cannot send, post, publish or pay — and that is not a setting, a permission or a limit being worked around. There is no send verb available to an agent at all. The app can send mail from its own mail screen when you press the button; that path is deliberately unreachable from an agent and is not one of the tools published to them.

This is the same promise the developer page makes about your main branch, applied to your outbox. In fact it is the stronger half: the code gate is a rule in a prompt, which a confused agent could ignore. This one is structural. There is nothing to disobey.

So a drafted reply arrives written, in your voice, with the context attached — and nothing goes out with your name on it that you did not read first.

The three agents, and what each one hands you

All three are ready-made prompts in the app, tuned and shipped rather than described. You can read every word of each one before you switch it on.

  • Inbox triage — your mail, read before you open it

    Every thirty minutes in business hours, an agent reads the inbox, sorts what needs you from what does not, and drafts the obvious answers. Connect a mail account over IMAP; the password lives in the macOS Keychain, never in a config file.

    Read-only access to your mailbox. It drafts; you send.

  • Renewal reminder — the things that expire quietly

    Daily, first thing: contracts, subscriptions, domains, certificates, licences. Anything with a date on it that you would have renegotiated if you had seen it coming. It files what it finds with the source link attached, so the decision is one click from the evidence.

  • Overnight backlog — the code half, unchanged

    Exactly the agent a developer gets: one item a night, in its own worktree, tested, left uncommitted, with an explanation beside each hunk. You are a developer too, and this pack does not pretend otherwise.

    By morning there is a finished, uncommitted diff with an explanation beside every hunk.

  • A run that fires when the right mail arrives

    A job can be triggered by mail matching a folder, a sender or a subject rather than only by a clock. That is a genuine event, and for this kind of work it is more natural than a schedule.

  • It asks before anything that costs or commits

    Spending money, sending anything as you, deploying, deleting — each blocks the agent's own process on a gate until you answer. A timeout, a refusal and an unreachable app all mean no. For a company of one, that is not bureaucracy; it is the only review there is.

  • Answer from your phone

    Through a bot you create, in a chat you own. No inbound port on your Mac and no third-party service holding your data — just a message asking which of two options you want, and the agent resuming when you tap one.

And nineteen more, already written

The three above are where you start. The template picker also ships a finance summary, an expense watch that hunts zombie subscriptions and price hikes, a plain-English first read of any contract that lands, a weekly strategy rollup against your own goals, and an end-of-day digest of everything that happened.

All of them are read-and-draft agents pointed at a scratch folder rather than a repo, and all of them end the same way: something written, waiting for you.

The questions worth asking first

How much time does the approve-and-send step actually cost me?

Read the draft, decide, paste, send. Call it twenty seconds each once you trust the drafting.

At three or four decisions a morning that is nothing against the reading it replaces. At forty it is a real chunk of your day, and you should point this at the highest-value slice of your inbox rather than all of it. That arithmetic is worth doing before you buy anything, which is why it is on this page rather than in a support thread. The app's own price belongs in the same sum: $89, once, against a month of mornings that start at five minutes of approving instead of two hours of finding out.

Does it connect to my payment processor or my accounting tool?

Nothing is bundled and nothing is endorsed — no payment, accounting, helpdesk or analytics integration ships with the app. If you already run an MCP server for one of those, agents can call it through the bridge. Otherwise the realistic path is dropping an export into a folder the agent reads, which is less impressive and does actually work.

Which agent CLI does this use?

Whichever you already pay for. Claude Code, GitHub Copilot CLI and agy (Antigravity) are all supported, and each is launched with the right flags for an unattended run — they differ, and normalising that difference is part of what the app does. You can also schedule a plain shell job for anything that is a script rather than an agent.

AgentCron contains no model and no subscription of its own.

Is my mail safe?

It never leaves your Mac except to your own mail server. There is no AgentCron backend to send it to. Passwords are in the Keychain; the connection is between your machine and your provider. The privacy page lists every destination the app can reach and how to verify that list yourself in about two minutes.

Does it need my Mac awake?

Yes. The scheduler runs inside the app, so a sleeping Mac fires nothing and there is no catch-up for a run that was missed. Settings has a keep-awake window — AgentCron holds the Mac awake between the hours you name, lid open — which is what makes an overnight agent real on a laptop. It is the limitation most likely to annoy you, which is why it is on this page rather than in a footnote.

See release status See pricing

Not released yet. It will need macOS 12.0 or later on Apple Silicon — and the download page says exactly what is still missing.