Skip to content

The night shift starts when you leave

Automate your AI coding agents. Then clock out.

Foreman is a free, open-source desktop app for macOS and Windows that runs AI coding agents in parallel. Claude Code, Codex, Cursor and OpenCode each work in their own git worktree on their own branch, while you review diffs, answer questions and open pull requests from one window.

Put prompts and flows on a schedule, let an on-call crew answer failing CI, and read what happened over coffee.

  • macOS & Windows
  • Apache-2.0
  • No account
  • v1.2.4

Try it · repo atlas · Europe/London

The shift board

Four schedules, an on-call crew and one laptop. Set the shift, press Clock out, and the night plays in nine seconds. Times are tonight, Tuesday into Wednesday.

TimesheetTue 17 → Wed 18 NovNow: 23:40, clocking out

You at the keyboard till 23:40

at the keyboardclocked outback

nightly-triage

.foreman/templates

nightly-triage, 00:30 to 00:52: planneddaily · max 30 min

keep-branch-fresh

built-in flow, experimental

keep-branch-fresh, 02:00 to 03:06: planneddaily · bounded by its loops

weekly-deps

.foreman/templates

weekly-deps, 03:00 to 03:24: plannedweekly, wed · max 30 min

release-notes

personal template

monthly · next Tue 1 Dec

On call

kyoto, osaka's PR #42, the reviewer

kyoto resumes, 00:00 to 00:18: plannedauto-fix #42, 01:12 to 01:31: planned01:10 CI failed on #4201:50 main moved01:44 CI green on #42

The laptop

its own sleep setting: 3 idle hours

sleep 02:40z z zThe computer slept from 02:40 to 07:00.
On call opt in

Holds off sleep only while an agent is working. It never wakes a computer that is already asleep.

Night log what Foreman wrote down
  1. kyoto stopped at the usage limit.
  2. You clocked out. The crew is on.
  3. The usage limit reset. kyoto resumed by itself.
  4. kyoto finished: +64 −9 on add-rate-limit · checks passed.
  5. nightly-triage started in a fresh worktree.
  6. nightly-triage finished: +46 −0 on triage-2026-11-18 · checks passed.
  7. CI failed on #42 fix-login-redirect.
  8. Auto-fix sent the failing logs to osaka, tab 1. Fix 1 of 2 for this commit.
  9. osaka pushed a fix: +7 −2.
  10. CI passed on #42.
  11. main moved. osaka still merges cleanly. Nothing rebased.
  12. keep-branch-fresh started in a fresh worktree.
  13. keep-branch-fresh was cut short when the computer slept.
  14. The computer went to sleep (its own sleep setting).
  15. weekly-deps was due. The computer was asleep: skipped, not replayed.
  16. You opened the lid. Foreman catches up on what it can see.

Preview: if you clock out now

Morning report · Wed 18 Nov, 07:00

Foreman

1 thing needs you

Illustrative push alert from the phone web app. iOS and Android apps: coming soon.

  • 3 ran
  • 1 PR ready
  • 1 needs you
  • 1 skipped

Needs you

  • keep-branch-fresh was cut short

    It started at 02:00; the computer slept at 02:40, mid-run.

    Run now

Ready for review

  • #42 fix-login-redirect

    CI green since 01:44 · draft · ready for your review

    Open in GitHub

What ran

  • kyoto · your chat 00:00–00:18

    Resumed at the limit reset · add-rate-limit · +64 −9 · checks passed

  • nightly-triage 00:30–00:52

    Labelled 9 new issues, flagged 2 duplicates, wrote docs/triage/2026-11-18.md · triage-2026-11-18 · +46 −0 · checks passed · workspace kept

  • osaka · auto-fix CI 01:12–01:31

    fix-login-redirect · +7 −2 · fix 1 of 2 for the commit, 1 of 5 for the PR

Skipped

  • weekly-deps 03:00

    The computer was asleep. A run more than 5 minutes late is skipped, never replayed.

The fix: turn on Keep the computer awake while agents run (Settings → General). It holds off sleep while an agent works. Nothing wakes a computer that is already asleep.

Also

  • main moved at 01:50

    osaka still merges cleanly. A base-moved notice never rebases, merges or starts an agent.

  • release-notes

    Monthly on the 1st at 04:00. Next run: Tue 1 Dec.

The fine print, lit up

Do schedules run when Foreman is closed?

No: schedules run only while Foreman runs. That doesn't mean a window. Foreman can run headless, or start with no window at every login, so schedules and flows keep going while the app looks closed.

Clock Foreman in at login
foreman service install

# A launchd agent on macOS, a Task Scheduler task on Windows. No administrator needed.

Night shift rules

  1. Claude Code works nights.

    Unattended runs use Claude Code today. A question or an approval stops the run and leaves it for you.

  2. Clock in with Foreman.

    Schedules run only while Foreman runs. Headless, or as a service from login, counts.

  3. No wake-up calls.

    A sleeping computer isn't woken. Keep-awake only holds off sleep while an agent works.

  4. Late is skipped.

    A run more than 5 minutes late is skipped, never replayed. No stampede at 07:00.

  5. New text, new sign-off.

    Change a template, its setup or its checks and its schedules pause until you approve again.

  6. Two strikes and you rest.

    Two failed runs in a row pause the schedule. Two scheduled runs work at once, at most.

By order of the foreman · see docs/cli.md, foreman schedules add

Automations

How do you automate AI coding agents in Foreman?

Save a prompt as a template, then run it on demand or put it on a schedule: hourly, daily, weekly or monthly, at a time in your time zone. Every run gets a fresh workspace, its own git worktree on its own branch, so nothing lands on main without you.

The seven templates that ship 7

  • Explain code{{file or feature}}
  • Fix a bug{{what happens and what should happen}}
  • Fix an issue{{issue id}}
  • Write tests{{file or feature}}
  • Refactor{{file or feature}}{{goal}}
  • Review my changesno blanks
  • Set up this repono blanks

Blanks fill in order: foreman run fix-a-bug "the login page loops". A scheduled run can't stop to ask, so the night shift runs templates without {{blanks}}. Repository templates live in .foreman/templates and travel with the code.

Schedule slipNo. 0042

nightly-triage on atlas

Every
hour · day · chosen weekdays · a date each month
At
00:30 in Europe/London, or any IANA zone
Runs on
Claude Code, a Sonnet by default
Max
30 minutes (5 to 480)
After
keep · archive · delete if unchanged
Approved
its text, the repo's setup, checks and MCP servers
foreman schedules add nightly-triage --repo atlas --every daily --at 00:30 --tz Europe/London
The Automations library of templates and commands, with schedules beside it
Automations Templates as cards, schedules beside them. Top right, the app says it plainly: “Runs only while Foreman is open.”

On call

What runs on autopilot between your turns?

Four opt-in helpers keep watch: auto-fix for failing CI within a budget you approve, a read-only auto reviewer, checks after each turn and base-moved notices. A chat that hits a usage limit resumes by itself.

  • Auto-fix CI

    When CI fails on your pull request, Foreman sends the failing logs to the tab you chose, once per failure, while the workspace is idle. It answers failing CI only, never review comments.

    2 fixes a commit · 5 a PR · 20 min each

    foreman pr auto-fix osaka on --tab 1
  • Auto review

    A read-only reviewer, Claude Code or Codex, reads each finished diff once. Its findings are an opinion and reach an agent only when you click.

    6 a day per repo · 10 min each

    foreman repo auto-review on --reviewer codex gpt-5
  • Checks after each turn

    Runs the repository's own check commands when a turn ends and nothing else is working. Checks are commands, never an agent.

    per repository

  • Base-moved notices

    Fetches the base and asks git whether your branch still merges. One notice per base tip; it never rebases, merges or starts an agent.

    needs git 2.38+

  • Resume after a usage limit

    A chat that stops at a usage limit waits, then carries on by itself when the limit resets.

    on by default

Flows on the rota Experimental

Can a flow run on a schedule?

Yes. With flows switched on, a schedule can start a flow instead of a template. Keep a long-lived branch fresh is built for it: overnight it merges origin/main into your branch, fixes what fails, then stops at a gate so you decide whether the merge is pushed.

Write a way of working down once as a graph of agent steps: loops that run until the checks pass or a review comes back clean, and gates where you decide. Turn them on in Settings → Experimental → Flows.

  • fix-until-green
  • research-then-implement
  • test-first
  • issue-to-draft-pr
  • second-opinion
  • two-reviewers
  • keep-branch-fresh
  • flaky-test-triage
  • document-what-changed
  • worth-building
  • idea-to-feature
A flow's graph of steps, loops and a gate, with the assistant under it
Flow Steps, a loop until the checks pass, and a gate where you decide.

The rest of the site

What else does Foreman do by day?

The night shift is one crew on a bigger site: parallel agents in their own worktrees, AI employees, control from your phone, integrations and remote sandboxes.

Remote control

Run it from anywhere

Pair your phone with a QR code to approve, steer and comment on diffs, live. Away from home it connects over your own Tailscale network, with no Foreman relay. A browser on another computer can open the whole app.

The phone reads terminals but can't type into them, and same-Wi-Fi pairing is direct. Access levels: Watch, Operate, Full control.

Today: pair your phone with a QR code and add the phone view to your Home Screen.

Needs you 2

keep-branch-fresh · gate

Push the merge?

Send backContinue

osaka · #42

CI green. Ready for review.

Illustrative

The day shift

One worktree per agent

Every workspace is its own git worktree on its own branch. Agents never step on each other, and your main checkout stays clean.

Tiles shows several workspaces side by side so you can step in where you are needed. Built for 10 to 12 agents running at once.

Standing crew

AI employees

Give an organization AI employees with a role in plain words, a hard hat and access you approve. They remember through a git memory repository.

Employees start on their own schedules and events too: a merged pull request, a Linear assignment, the base branch moving.

Connections

GitHub, GitLab, Linear, Slack, Langfuse

Pull requests, CI and reviews on GitHub and GitLab, work from Linear, notifications to Slack, traces from Langfuse. You pick what agents may touch: nothing, read, or read and write.

Slack is outbound only: notifications and posts.

Off-site · experimental

Remote sandboxes: E2B, Daytona, Vercel, Docker

Run a workspace's worktree and agent on your own E2B, Daytona, Vercel Sandbox or Docker. Provider keys never enter the sandbox, and turns keep going after you close the app.

Claude Code, Codex and OpenCode; not Cursor. A sandbox needs a GitHub remote.

Other sites

How is Foreman different from Conductor and T3 Code?

Foreman is open source under Apache-2.0 and runs on Windows as well as macOS, while Conductor is a closed-source Mac app. Only Foreman lists scheduled automations, auto-fix CI and flows.

A dash means not listed on their site

Foreman compared with Conductor and T3 Code
FeatureForemanConductorT3 Code
Open sourceApache-2.0NoMIT
macOSYesYesYes
WindowsYesMac onlyYes
One git worktree per agentYesYesYes
AgentsClaude Code, Codex, Cursor, OpenCode, any ACP agentClaude Code, Codex, CursorClaude Code, Codex and more
Flows that loop until checks pass, with human gatesYes——
AI employees with git-backed memoryYes——
Scheduled automations and auto-fix CIYes——
Your own cloud sandboxes (E2B, Daytona, Vercel, Docker)Experimental——
Scriptable CLI, --json on every commandYes——
Approve from your phoneYes—Yes
No account, no telemetry uploadYes——
From each project's own site as of November 2026. “—” means it isn't listed there, not that it's missing. Spot a mistake? Tell us.

First shift

Get started in a minute

Download the app, point it at a git repository, and give each task its own workspace. Schedule the ones that repeat.

  1. Download Foreman

    Free for macOS (Apple Silicon) and Windows. You need Git on your PATH.

    Download Foreman
  2. Add a repo, sign in an agent

    Pick any git repository. Sign in Claude Code, Codex, Cursor or OpenCode with your own login or key. Schedules use Claude Code.

  3. Create workspaces, then schedule

    One workspace per task. When a prompt repeats, save it as a template and press Schedule… on its page.

Or from a terminal: run a template now, schedule it later
foreman run fix-a-bug "the login page loops after sign-in" --repo atlas --wait

# Start a workspace from a template and wait for the turn. A schedule prints what it will approve and asks first. CLI reference

Toolbox talk

Questions developers ask

Something missing? Open an issue or read the guide.

What is Foreman?

Foreman is a free, open-source desktop app for macOS and Windows that runs AI coding agents in parallel. Claude Code, Codex, Cursor and OpenCode each work in their own git worktree on their own branch, while you review diffs, answer questions and open pull requests from one window.

Is Foreman free and open source?

Yes. Foreman is free and licensed under Apache-2.0, with the full source on GitHub at github.com/elite-build/foreman. There is no paid tier, no account and no Foreman cloud service.

Which coding agents does Foreman support?

Foreman runs Claude Code, Codex, Cursor and OpenCode, plus any agent that speaks the Agent Client Protocol (ACP). Each agent signs in with your own subscription or API key, set in Settings → Agents.

How do I run multiple Claude Code agents in parallel?

Add a git repository to Foreman and create one workspace per task. Each workspace is a separate git worktree on its own branch, so several Claude Code sessions edit the same repository at once without conflicts. Tiles shows them side by side.

Why does each agent need its own git worktree?

A git worktree is a second checkout of the same repository on a different branch. Giving each agent its own worktree means agents never overwrite each other's files, every change lands on a reviewable branch, and your main checkout stays untouched.

Does my code leave my computer?

Only through the agents you choose. Foreman runs locally with no account and no telemetry upload; your agents talk to their own model providers. In your own workspaces, an approval marks any command that leaves your computer before it runs, and AI employees hold Slack posts and Linear changes for you.

Do coding agents need my approval for every command?

Only if you want them to. Approvals are a setting: with Claude Code tool approvals on (Settings → Account), each command is spelled out part by part before it runs, anything that leaves your computer is marked, and Always allow shows exactly what it adds. Every agent still works in its own worktree, and checkpoints rewind any turn.

Which operating systems does Foreman run on?

Foreman runs on macOS (Apple Silicon, signed and notarized) and Windows. You need Git on your PATH and a login or API key for at least one agent.

Can Foreman run coding agents on a schedule?

Yes. Any saved prompt template can run hourly, daily, weekly or monthly in your time zone, and opt-in auto-fix sends failing CI logs to your agent. Scheduled runs use Claude Code and happen while Foreman is running, which can be without a window or from login.

Can I control Foreman from my phone?

Yes. Pair a phone with a QR code to answer approvals, steer agents, comment on diffs and read terminals, live. On the same Wi-Fi it connects directly, and anywhere else through your own Tailscale network. Native iOS and Android apps are coming soon.

Can coding agents run in a cloud sandbox like E2B?

Yes, as an experiment. A workspace's worktree and agent can run in a sandbox on your own E2B, Daytona, Vercel Sandbox or Docker account, for Claude Code, Codex and OpenCode. Provider keys stay on your computer and never go into the sandbox.

Which integrations does Foreman have?

GitHub (pull requests, merges, CI and reviews), GitLab, Linear, Slack and Langfuse, plus MCP. Each connection has a default per service, each repository can pick its own, and you choose what agents may do with it: nothing, read, or read and write.

How is Foreman different from Conductor?

Foreman is open source under Apache-2.0 and runs on Windows as well as macOS, while Conductor is a closed-source Mac app. Foreman also drives OpenCode and any ACP agent, and adds flows that loop until checks pass, AI employees and a scriptable command line.

How is Foreman different from T3 Code?

Both are open-source GUIs for coding agents. Foreman adds flows (agent graphs that loop until tests pass, with human gates), AI employees with a git-backed memory, a foreman CLI where every command takes --json, phone approvals and remote sandboxes.

Can I script Foreman or use it in CI?

Yes. The foreman command starts workspaces, sends messages, answers approvals, opens and merges pull requests and runs flows. Every command takes --json and exit codes are the same for every command.

The crew works nights. The code is open all day.

Foreman is three weeks old and has shipped 31 releases since 1.0.0 on 27 October, in the open under Apache-2.0. Star it to hear about the next one, fork it, or pick up a first issue.