Scheduled Tasks
- Desktop, mobile and web
- All plans
What it does
A trigger is a prompt, an agent (or a whole team) and one question: what makes it run? Scheduled tasks answer with the clock (every N minutes, hourly, daily at one or several times, weekly, monthly), with your AI subscription's quota window (when the quota comes back, or at the tail of the window before it resets), or with nothing at all: an On demand trigger has no clock and no endpoint and runs only when you press its button, which is what a job you repeat but never on a rhythm looks like (build the app, cut a release). When it fires, AgentsRoom starts an agent in the project, hands it your prompt as its first message, notifies you, and archives the run so it can be read later from any machine of the account. No cron expression is ever typed.
Video (1 min 16 s): Run now instead, shown in the real app. In Automations, the day's timeline shows a review agent planned every day at 22:00; hovering its card and clicking "Run now instead" starts the agent at once, its first message says it replaces the 22:00 run, the card turns to "Skipped this time", and tomorrow's 22:00 is still planned.
Where to find it
- Inside a project, the bar welded on top of the workspace dock (since 2026-10-28; before, tiles and "When it runs" chips at the end of the agents list, and a Triggers button in the dock): New agent opens a menu in two groups. With you: Agent, Team, Throwaway agent, Brief ready to launch (on demand). Without you ("it starts on its own"): At a set time, When your quota comes back, Started from outside (webhook), After a delivery (on ship). Each trigger row has a plain title and a one-line description (hover it for a concrete example; the phone shows the example under the row), and opens the Triggers window filtered on that kind; its + creates one of that kind.
- Same bar, right side: Automations (with the number of triggers, an amber dot while a run works, a green one when a run finished unread) opens the Triggers window on every kind.
- The agent advisor ("Help me choose an agent", also the first line of the New agent menu): when the task you describe is recurring or event-driven ("an agent at regular intervals", "review every pull request", "when my quota comes back"), it first shows This looks like an automation with the matching mode and Set up the automation, which opens the Triggers editor on a draft of that mode briefed with your sentence.
- An agent's right-click menu: Do this task without me… opens the Triggers editor on a NEW trigger already filled with that agent's role, provider, model, account and first message; pick when it runs and save.
- With no trigger yet, the empty panel shows five cards: On a schedule, On your quota, On an event, When work ships (since 2026-09-30, see Webhook Triggers), On demand; clicking one opens the editor preset on it.
- The filter above the list shows, on each kind of trigger, how many triggers it holds (counted with the current search, so the number is what a click would show). As soon as one scheduled trigger exists, the panel opens on Scheduled, even when webhooks or other kinds sit next to it (since 2026-11-10). Without a scheduled trigger, a project whose triggers are all of one kind opens straight on that kind, otherwise on all of them. A kind you picked yourself is never moved. Same on the phone.
- Scheduled opens on a Timeline of the day by default (a Timeline / List switch sits next to the filters; the choice is saved on your account, so every computer and the phone open on the same view). The timeline shows the week on top (each day with how many runs it holds, weekly and monthly triggers on their own day), the day's tally (done, running, failed or missed, to come) with who runs next and when, the triggers firing every few minutes or hourly in an All day long band, then the day on a vertical axis: one card per run at its exact time, a dot on the spine coloured by its state, the elapsed part of the day filled in and a "now" line. A card says Ran at 09:02, Running (a run working here or on another machine of your account), Failed to start, Missed, Did not run or a countdown (in 42 min); paused triggers stay visible, faded.
- Hover a card on the timeline to open its detail card: everything the List row shows (rhythm, who runs it, the machines it may fire on, provider, model and reasoning effort, the team's crew), where this run stands, the next and last run, the start of the prompt, and Open console (while that run's console is alive; Open console on <machine> when it runs elsewhere), Edit, Run now, Run now instead (only on the trigger's next run, when the trigger is on and that run is not already skipped), then Pause: This run only skips just this occurrence (the card turns to Skipped this time, every other run fires normally, Don't skip brings it back) and All runs switches the trigger off, like the List switch (Resume turns it back on). Desktop only: the phone has no hover, a tap opens the editor and the pause button at the end of each card offers the same two choices.
- On the timeline, drag a card up or down to move its trigger to another time (15-minute steps, Alt + arrows from the keyboard); a bar confirms the next run and offers Undo. Hover an empty half-hour to get a ghost card: Every day or <weekday> only opens the editor already set to that time. On the phone: tap an empty slot, then a choice; long-press a card, then drag. Moving a slot earlier than now does not run it today: the next run is tomorrow.
- The List shows each trigger's next run (or its last one, for a webhook and an on-demand trigger), enable switch, what it runs (a single agent's provider and model; for a team, one chip per member with its provider logo and model, on the phone too), Run now, Run history, and a filter (All, Scheduled, Quota, Webhooks, On ship, On demand), each carrying the icon of its kind. Filtering on a kind you have none of says which one is missing (No scheduled agent yet, No quota agent yet, No webhook agent yet, No on-ship agent yet, No on-demand agent yet) instead of a blank list. The Command Palette also searches triggers.
- Select (next to New trigger, in the List) turns on selection: tick the triggers to change, then use the bar at the bottom: Enable, Disable, Delete (asks once), or Edit together, which opens the usual editor on the first selected trigger with a notice "Editing N triggers at once". On save (Apply to N triggers), only the fields you changed are written to every selected trigger; everything else (their day, their name, their own prompt if you left it alone) keeps each trigger's own value. Edit together needs at least two triggers of the same type: Scheduled, Quota or On demand (webhook and on-ship triggers each own their address, so they are enabled, disabled or deleted together but not edited together). Select all / Clear, Done to leave.
How to use it
- New trigger. Fill Name (may contain
{{date}},{{time}},{{datetime}}, resolved at each run) and Prompt. - Under Trigger, pick On demand (nothing else to set: it waits for you), or Scheduled and a Frequency (Minutes, Hourly, Daily, Weekly, Monthly) with its time fields; on Weekly, tick one or several Days (Weekdays picks Monday to Friday in one click), and on Daily and Weekly, Add a time fires the task at several times of day (up to six). Or pick Quota: What fires it = When quota returns or Before reset, then the Quota window (5-hour, Daily, Weekly, Monthly), Run before reset (30 min to 4 h) and Minimum used (Any usage, 50%, 80%, 90%).
- Under "Who runs it", choose One agent (Role from the same catalogue as Add Agent, Provider, Model, Reasoning effort), A team (the prompt goes to the team's first step), or Existing project agent ("Send the prompt to a saved agent in this project, keeping its conversation."): pick it under Choose an agent; delivery waits while that agent works or needs your answer, and repeated reminders are combined into one pending message. Reopen the agent if its console is closed is optional and off by default.
- Advanced carries what an unattended run needs: Permission mode (Inherited, Autonomous, Manual), account, Browser access, Always memorize, Default skills, System prompt, CLI options and environment variables. Empty means inherited.
- Machines: All machines or chosen ones, so two computers with the project open do not both fire the same run.
- Options: Enabled, Notify when it starts, Close after inactivity (with a delay), Catch up if missed, Wake the machine (Inherit, Wake, Do not wake). Save.
Run now fires immediately and opens the agent's tab, as an extra run: the planned ones still fire. Run now instead (timeline, detail card of the next run) starts the agent at once in place of that next run: the slot turns to Skipped this time, the following runs stay planned, and the agent's first message starts with "Note from AgentsRoom: you are a scheduled agent, but this time you were launched by hand, in place of your run planned at 10:00 PM, which will not take place.", so a routine written for one run a night does the work instead of stopping. On the phone, Run on which machine? has the switch Replace the next run (HH:MM) instead of adding one. Opening a run's console (Run now, or opening a run from the history) closes the triggers window to make room; closing or hiding that agent afterwards brings the window back where you were, same filter, same trigger unfolded (switching to another agent cancels the return). Right-clicking it (or long-pressing, or pressing the down arrow) offers the other machines of your account that hold this project, so a build can be fired on the machine that has the toolchain from the laptop you happen to be sitting at; the run is stamped and archived over there and shows up in the same history. The mobile app has the same button, and always asks which machine, since a phone runs nothing itself. In Run history, Open reopens a run still alive on this machine or attaches to it on another machine of your account, Read opens the archived transcript of any past run, including runs from another machine. Only a run you reopened that way gets a terminal tab and comes back after a restart: a run working in the background does not, since its brief is delivered once, when it fires.
History (editor, existing trigger): every save that changes a trigger (name, prompt, schedule or webhook, agent or team, provider, model, options) keeps the previous configuration; runs and the last run date are not versions. The selected version is compared line by line with the draft (name, prompt, then one line per option), and Restore this version loads it in the editor as unsaved edits: save to keep it, the current state stays in the history.
Settings
triggerScheduleView(account, the Timeline / List switch of Triggers > Scheduled, desktop and phone):timeline(default) orlist. Synced to the account entry by entry, so changing it on the phone is never undone by a computer.wakeForTriggers(global, Settings > Terminal > Wake the machine for scheduled triggers; project override in Project settings; per-trigger override in the editor): arm the operating system's wake timer for the next run. Off by default. macOS and Linux need a one-time administrator authorization shown in the same place, revocable there.autoLaunchAgentsAtStartup(global, Settings > Terminal): allows runs missed while the app was closed to be caught up at launch; off means nothing starts until you ask.keepAwakeWhileWorking(global, Settings > Terminal): keeps the machine from sleeping while the run's agent works.
Agent tools (MCP)
triggers_list,triggers_get: read the project's triggers and their runs.triggers_save: create or update a trigger (schedule, quota, webhook or manual mode, agent orteamId). A weekly trigger takesweekdays(e.g.[1,2,3,4,5]for Monday to Friday, 0 = Sunday) or a singleweekday.skippedOccurrences(ISO datetimes of the slots, local time) skips those runs once without disabling the trigger;[]clears it,enabled: falsestops every run.triggers_delete: remove one.triggers_run: fire a trigger now; returns the agent id, or the run id for a team.
Providers
All providers: the run goes through the same launch path as a manual agent. The quota frequency depends on AgentsRoom being able to name an exact reset moment for the account: from a structured source (the Claude quota API, the Codex usage API or its last session snapshot, the Grok, Cursor, Kimi and OpenCode usage APIs, Antigravity's agy /usage) or from the reset sentence printed in the CLI's own usage panel, which is now read back into a date for Kimi and Grok too. When nothing can be resolved the editor says so and offers no window.
Mobile
Present, timeline included: same week strip, tally, axis and card states; the pause button of a card offers This run only / All runs; tap an empty slot to create, long-press then drag a card to change its time, Undo in the bar. The phone runs nothing, but a run working on a machine of your account reads Running, and the terminal button of its card (or Open console in the history) attaches to it. The phone reaches the project's triggers from the workflow glyph in the project header, with the same list, the same four kinds and their icons, the same filters and named empty states, the same editor, history, machine scope and agent / team choice, using the same words. Two differences, both deliberate: Run now always asks Run on which machine? (and says "None of your machines is online right now." when none is), because a phone runs nothing itself; and a quota trigger shows Waiting for a quota reading until a desktop has read the provider's usage window. The editor has the same Version history; the version is loaded into the phone's editor and saved like any other edit. The weekly Days pills (with Weekdays) and the selection are there too: the list icon in the header or a long press on a trigger starts the selection, a tap ticks a trigger, and the bar at the bottom offers Enable, Disable, Edit together and Delete.
Limits
- Runs fire while AgentsRoom is open on an allowed machine. A missed run is caught up on the next launch unless Catch up if missed is off, in which case a slot missed by more than ten minutes is recorded on the task row as Last run: Missed. Whenever a run is missed or fails to start, AgentsRoom also sends a desktop notification and a phone push (one grouped notice when several are missed at once, for example after a night with the app closed), even when Notify when it starts is off.
- Wake timers: a laptop on battery may ignore them; keep it plugged in. A closed lid on power works. On Windows, wake timers must be allowed in the power plan.
- Quota precision is that of the usage scan (about 10 minutes through the Claude OAuth API, 30 through the usage terminal), so a quota run starts shortly after the boundary.
- Team runs started by a trigger share the project root (no worktree); every step of a run is archived together in one row of the trigger's history. Close after inactivity applies to them too since 1.184.0: the run closes once its consoles have been idle for the delay, even while their tabs are on screen (before, a team run whose consoles were visible never closed on its own). Without that option, a finished team run waits for Finish run (or
team_finalize_run). A team run closed this way that has a ticket sends it to In Review when the board has that column (since 2026-10-22), since nobody checked the work. Since 1.208.0 a team run closed this way also leaves the project's agent list, as if you had clicked the ✕ of its box: its transcript stays readable in the trigger's history. A team run of that trigger you finished yourself leaves the list once it has been finished for the same delay. Before, every closed run stayed in the list as Done or Cancelled · Closed after inactivity until you clicked ✕ on it. - Close after inactivity now survives a restart or a crash: a background run's console is not brought back on launch, so it can no longer escape the timer and stay open until you close it by hand.
- Close after inactivity judges a run on what its agent does. When the CLI draws a full-screen interface whose state AgentsRoom cannot read (Antigravity, Mistral Vibe, Copilot, or Claude Code in fullscreen mode without its hooks), the run counts as inactive once its terminal has printed nothing for the delay (since 1.199.0; before, such runs never closed on their own). A run you are looking at (shown in the project you have open, window in front, you at the keyboard) is never closed, and the delay starts again once you leave it; a run sitting in a tab you are not looking at is closed like any other, its transcript archived first.
- A trigger run's console is not woken by your agents' project mail (a broadcast, or a message sent to every live member): only you, AgentsRoom itself, and replies in a conversation the run started reach it. Before, every such message gave each open run a paid turn and restarted its Close after inactivity timer, so runs never closed. The message stays in the run's mailbox.
- Every live run uses one of your plan's parallel agents. When they are all taken, a new run waits and starts by itself as soon as one is freed: Close after inactivity is what frees them on an unattended machine.
- The run history of a trigger keeps its five most recent runs on the machine; when an older one falls out, its console is closed with it. The transcript of any past run stays readable with Read.
- Times use the machine's local timezone.
- Run now instead exists only on the next run of an enabled trigger with a schedule (daily, weekly, monthly cards of the timeline): triggers every few minutes or hourly have no card (they sit in the All day long band), and a later run cannot be replaced, only skipped.
Common questions
- Can I see which scheduled agents are running, and open their console, even on another computer? Yes: in Automations > Scheduled, a run still working reads Running on the timeline, and any run whose console is alive has a terminal button on its card. Open console on <machine> in its detail card attaches to it over there, it never starts a second agent. Triggers firing every few minutes or hourly have the same button on their chip in the All day long band. On the phone too, if it is signed in to your account.
- A run on another computer waits for me. Do I need the phone? No: any desktop signed in to the account lists the run while its console is alive, marked Trigger, with the name of its machine. Click it to answer there. See Remote Fleet.
- Will I know if a trigger did not run? Yes. A missed run (the app was closed, the computer asleep, or the trigger does not catch up) or a run that failed to start sends a desktop notification and a push to your phone, and is written in the trigger's history as Missed or Failed to start. Several missed at once arrive as one notice listing them. With Start agents automatically at launch off (Settings > Terminal), runs due while the app was closed are not started late: you are told they were missed.
- What if the CLI account of a trigger is signed out on that machine? The run is held instead of started into a "Please run /login" refusal: you get a Waiting notification on the desktop and the phone, and the run starts by itself within a minute of signing in again (
/loginin any agent of that account,claude auth loginin a terminal tab, or from the phone's terminal). On a machine pool, another signed-in machine takes it. Details: CLI Doctor. - Does it run when AgentsRoom is closed? No. It is an in-app scheduler with catch-up, not a server-side cron. Turn on Wake the machine to have the computer wake for the run.
- Will the agent get stuck on a permission prompt at 3 AM? Not if the task's Permission mode is Autonomous; your manual agents on the same provider keep asking.
- Can I run an agent when my Claude quota resets? Yes, the Quota trigger with When quota returns.
- Can I stop my scheduled agents when my quota is almost gone? Yes, with a quota rule: New rule with the action Block a launch and Which launches set to Scheduled agents blocks scheduled launches of that provider once the threshold is reached (for example 99 %). A blocked trigger starts no agent and sends a notification instead. See Quota Rules.
- Does a trigger that runs a team respect quota rules? Yes, since 2026-09-23. The run is judged on the provider and account of the team's first agent; if a rule blocks it, the run does not start (or stops with "Quota rule "X": run stopped"), and every agent of the run keeps the trigger's origin (scheduled or webhook) for the rule's Which launches scope.
- Can a quota threshold start a trigger? Yes: a quota rule with Run a trigger starts an existing trigger once per quota window.
- Can a trigger send its prompt to an agent I already have, instead of a new one? Yes: Existing project agent, then a saved agent of the project. The agent keeps its conversation, so a trigger every 15 minutes can ask it to check the backlog. A closed console is reopened only if you tick Reopen the agent if its console is closed, and its previous conversation is resumed only when the provider and your restore settings allow it. Same choice on the phone.
- I asked my coordinator to watch my PRs overnight and found every agent idle in the morning. Why? An agent only works during a turn. Once its answer ends, nothing wakes it except a message, a trigger aimed at it, or a background task of its own CLI: a PR update, a CI result or a teammate that stops without reporting does not reach it. Give it a heartbeat: a schedule trigger with Existing project agent, the coordinator picked under Choose an agent, every 15 to 30 minutes, and a prompt that says what to check (open PRs, its delegations) and when to stop. Keep the desktop app open and the computer awake. An agent asked to keep watching while you are away now says so before you leave and offers to set that trigger up; a reminder waits while the agent needs your answer, so it should not end its turn on a question. Teammates in the chain follow the same rule: a closed console or a turn ended on a question keeps their mail waiting for you, unless you turn on A message may start the recipient or Deliver mail to an agent waiting for an answer (Settings > Agents & AI > Agent behavior, both off by default).
- Can I pause a scheduled agent just for today? Yes: hover its card on the timeline, then Pause > This run only (on the phone, the pause button at the end of the card). Only that run is skipped; the trigger stays on and the next one fires as usual. All runs pauses it until you resume it. A skip is saved on your account, so the machine that runs the trigger honours it even if you set it from another computer or the phone (both should be on the same time zone, since the slot is a local time).
- How do I run my nightly agent now without it running again tonight? Use Run now instead: Automations > Scheduled > Timeline, hover the card of its next run (for example 22:00), click Run now instead. The agent starts right away and is told it replaces that run, the 22:00 card turns to Skipped this time, and tomorrow is still planned. Only the next run can be replaced; Run now alone adds a run. On the phone, the same choice is the Replace the next run switch when you launch it on one of your computers. Video: https://youtu.be/sUoEtmgitvc
- Is there a video that shows it? Yes: Run now instead, shown in the real app (1 min 16 s), from the day's timeline to the skipped 22:00 and tomorrow's run.
- Can I see which agents run today, and when? Yes: Triggers > Scheduled opens on the Timeline of today, with the next run and what already ran. The week strip shows the other days, weekly and monthly triggers included.
- How do I change the time of a daily agent quickly? Drag its card on the timeline; Undo puts it back.
- Can I go back to the list? Yes, List next to the filters. The choice follows your account.
- Can a task run twice a day? Yes, Add a time on a Daily or Weekly task.
- Can a weekly trigger run Monday to Friday only? Yes, since 2026-10-14: on Weekly, tick the days you want, or click Weekdays. Before, it took one trigger per day. Older versions of the app still fire such a trigger on its first day only, so update every computer of the account.
- I have the same trigger several times (one per day). Can I change them all at once? Yes: Select, tick them, Edit together, change what you need, Apply. Only what you changed is applied, so each trigger keeps its own day. You can also merge them into one weekly trigger on several days and delete the others.
- One scheduled slot started several agents at once. Is that expected? No: each slot runs once. Fixed on 2026-10-07: a slot is now closed the moment it fires, so slow network or app responses (or waking from sleep) can no longer start it again on the next check. Run now and webhook events still start a run every time.
- Where are tasks stored? Per project, synced to your account, with a local cache.
- My trigger runs stay open although Close after inactivity is on. Why? Before 1.199.0, a run on a full-screen CLI (Antigravity, for example) never closed on its own, and neither did a run sitting in a terminal tab, even one nobody looked at. Update the app: such a run now closes once it has done nothing for the delay. Only the run you are actually looking at stays open. An update only applies once AgentsRoom is really quit (Quit, or Update ready in the footer) and reopened: an app left open for weeks keeps running its old version. If an agent checks the runs with
agents_list_live: a run still listed asidleorbusyhas a console that is really open, and since 2026-10-16 a closed run leaves that list (before, it stayed there asofflineuntil the trigger's history dropped it). - A run finished hours ago is still open although Close after inactivity is on. Fixed on 2026-11-10: on a project where another trigger fired every hour or so, a finished run's idle prompt could look busy for a few seconds and restart its timer, so it never reached the delay. The run now counts as working only while its CLI really runs a turn. Update the app.
- My scheduled team runs still show in the agent list as Done although Close after inactivity is on. Why? Before 1.208.0, closing a team run ended it but kept its row in the list until you clicked ✕. Update the app: a team run closed after inactivity now leaves the list, and the rows already piled up go away by themselves within a minute (a run you finished by hand stays for the trigger's delay). Each run stays readable in the trigger's history. Without Close after inactivity on the trigger, a finished run keeps its box until you click ✕.
- I changed the prompt of a scheduled agent and the results got worse. Can I go back? Yes: open it, History, pick the earlier version, Restore this version, then save. The last 50 versions are kept, and a restore can itself be undone.
Related
- Webhook Triggers: the same panel, fired by an outside event.
- Message Queue: the one-shot version of quota timing.
- Quota Rules: block, redirect or adapt trigger launches when a quota is high.
- Keep the Machine Awake: what keeps the machine up while the run works.
- Agent Teams: point a trigger at a team.
- Agent Notifications: the start notification and mobile push.
- Prompt References: point a trigger at a library prompt instead of pasting it.