2026-08-27 · Guide

How to Stop a Grok Bot Roster From Burning the Weekly Pool

Blank Tuesday standup is the receipt you get when Monday's five-minute inbox bot spent the weekly pool, because the product still has no Bot-specific spend cap and will not name which card did it.

To stop grok bot overspending this hour you do four things in order. Pause every routine. List the 50-cap owners. Delete the five-minute inbox bot. Refuse retries. Charters and a twice-daily replacement clock are later hours.

This page is not the quota exceeded banner page, which decodes the message. It is not the spend-cap policy page, which writes ceilings into charters because the product has no per-bot cap. It is not how to schedule a Grok Bot routine, which attaches a new clock after the fire is out. Stay here until the poller is gone and the other two cards are still paused.

All bots share one persistent cloud computer assigned to the user, not to a bot (computer and apps). The weekly usage pool follows the same grain. There is no Grok Bot-specific spend cap, only the account-level On-demand monthly limit (teams and enterprises). Plans that include Grok Bot come with a weekly usage allowance. Past that pool, work is billed on demand from model and token cost (Grok Bot FAQ). No published page prints the allowance as dollars, credits, or runs. This page will not invent one.

Pause every routine from the phone before you hunt a spend slider

There is no per-bot slider labelled spend cap. The account-level On-demand monthly limit exists, in Grok Bot Settings or under Spending on cursor.com/dashboard, but a bot already working can finish past it, and searching threads for it is how the five-minute clock gets another dozen fires. Pause is the control you have in the first two minutes.

The phone can pause and resume, open run history, and delete a routine. It cannot edit or test. Those need a desktop (mobile). If you are away from that desk, pause is the whole hour.

Pause every standing routine, including the quiet ones. The bot that looks broken is rarely the bot that emptied the pool. Inbox Triage on five minutes can take down Chief of Staff Briefing and Standup Scribe without those two ever looking busy. Lead Scout and Churn Watch on daily clocks can look innocent in the same hour. Individual accounts and self-serve Teams still have no audit view of Bot actions; Enterprise has audit logs and Action Recording. Treat the roster as one tank.

Do not delete in order to pause. Deletion is a later move on one card, at a desk, after you have a written owner list. Leave the roster names in place until you know which name is the poller.

There are desktop apps for macOS, Windows and Linux, and companion apps for iPhone (iOS 18 or later) and Android (9 or later); the iOS app also runs on iPad (FAQ). Switching devices does not refill the week. See supported platforms.

Write bot name, routine name, paused at. Do not retry. Do not start a manual run just to see. Do not open teach-by-demonstration.

List the 50-cap owners so a quiet bot cannot hide a live clock

A routine assigns a workflow to one Bot. Max 50 routines per Bot. The app keeps the 20 most recent run records per routine. Deleting a Bot deletes its routines. Nothing is team-level (skills, routines, and automations).

That 50 is a per-bot cap, not a team calendar. There is no account-wide owner screen. The spend map you actually have is the list you write: each bot card, how many of its 50 slots are filled, and the tightest clock still attached. You cannot see cost. You can see clocks.

Open each bot at a desk. Copy the routine name, the schedule, and whether it is paused. Twenty run records is a cache, not a bill. Keep anything you still need in a file you own.

A bot with zero routines is not burning the pool on a clock. A manual run still spends, so do not test it. Rank clocks. Five minutes is the fire. If two jobs share one card, write that down: one delete would kill both. Inbox Triage must not also write the standup.

Delete the five-minute inbox bot after pause, not after a coarsen promise

Pause stops the next fire. Resume brings it back in one tap, including from iPhone. That is why the five-minute inbox bot has to leave the roster this hour, not on Friday when you swore you would coarsen it.

Delete only that card, and only at a desktop. If you are still on the phone, stay in pause. Do not resume. Do not create a replacement from the commute.

Copy the routine wording into a file you own first. Deleting a Bot deletes its routines and the 20 stored run records. If you want the charter later, duplicate the card, disable the copy's routine, then delete the original so the live five-minute clock cannot be resumed. Duplicate does not isolate cookies. Two enabled copies is two fires.

Hide is not delete. Hide does not pause. A hidden inbox bot on five minutes is still a five-minute inbox bot.

Deleting the poller is a spend move, not a teardown. It removes the name, the chat, and the routines. It does not sign out Gmail. Mail Cleanup Assistant can still open the same mailbox session. How to delete a Grok Bot without leaving logins behind is the revoke order. Do not revoke Gmail in this hour unless you already wanted it gone. Sessions live across bots (shared computer).

Do not delete Chief of Staff Briefing or Standup Scribe to save the pool. Those cards were not the five-minute loop. Killing them does not refill the week. Grok Bot and Gmail is the draft-only page. Rebuild mail later, twice a weekday, never every five minutes. Cadence sits in Grok Bot scheduling.

After the poller is gone, leave the other two paused. Tuesday's briefing and standup are work you do as a person, skip, or run once by hand after you confirm overflow in the product. The phone can resume. That is the trap at 07:28.

Walk three bots from a Monday enable to a silent Tuesday

Three bots. One clock at five minutes. Pool gone Tuesday. This is a shape, not a measurement. Nobody can tell you which minute an unpublished pool died.

You wanted Inbox Triage to feel live. Twice a day felt slow. You set every five minutes, labelled mail only, draft replies, never send. That charter is otherwise right. The clock is the mistake. Chief of Staff Briefing stayed on weekdays at 07:30. Standup Scribe stayed on weekdays at 09:00, DM only. You did not think of them as a roster. The weekly pool did.

Every five minutes is 12 fires an hour, 288 in a day if nights stay on. Monday 08:00 to Tuesday 07:30 is about 282 fires from the inbox card. The briefing ran once Monday. The standup ran once Monday. Two useful runs, then hundreds of polls against the same unpublished weekly pool. Do not read 282 as a price. There is no published price per run. Read it as a count.

The tell is a different bot going quiet on Tuesday. Then you retry that quiet bot, and the five-minute clock keeps firing. Twice a weekday would have been two inbox runs by Tuesday morning, not hundreds.

Treat every retry as a purchase you have not signed

A blank pack and a failed run invite the same click: try again. That click is how you turn a stop into a second bill.

A failed run still spent the attempt that failed. A retry spends another. If overflow is on, each retry is on-demand work from model and token cost. If overflow is off, each retry is a wasted loop. Clicking again is not how the week refills. A plan change is not a documented mid-week refill. If the product does not say it, this page will not say it either.

Confirm overflow in the product and on the vendor's current pricing page. Every paid Cursor plan (Pro, Pro+, Ultra) and every self-serve Cursor Teams seat include Grok Bot, an individual SuperGrok, SuperGrok Plus, SuperGrok Heavy or X Premium+ subscription can be linked, and a one-time trial exists (plans). Confirm names and prices on cursor.com/pricing and x.ai/pricing. None of them print a dollar figure for the included weekly allowance. There is no model picker, so you cannot limp through on a cheaper model.

An approval only covers the next proposed step. Work already done is not undone, and it is not refunded (approvals, security and privacy). Denying a send does not refill the pool. The 20-record window is not a receipt. Write the time and the wording you saw, then stop clicking.

Park this hour off the banner page and the spend-cap policy page

Keep these pages apart or you will do the wrong hour of work. This page is the bleed-stop. Grok Bot quota exceeded decodes the banner. The spend-cap page is the monthly limit and the charter ceilings you write after. How to schedule a Grok Bot routine is how a replacement clock gets a named owner. Grok Bot cost is how usage adds up while you design a roster.

If you searched stop grok bot overspending, stay until every routine is paused and the five-minute owner is gone.

Sort pause, delete, coarsen, and wait as four different moves

People mash these into one mood. They are four moves. Only two belong in this hour.

Pause without delete leaves a resume button on the poller. Delete without pause is how people smash the wrong card from a phone they cannot delete on. Coarsen without delete is a promise. Wait without pause is how overnight fires keep going. Least privilege and the safety checklist are later hours.

Paste a bleed-stop pack the remaining bots will actually obey

Pause is this hour's hand. The pack is how the next empty pool is quieter, and how a surviving bot does not retry its way onto on-demand. Paste it into a file you own, then into every remaining charter.

BLEED-STOP (this hour, you do this, not the bot)
Pause every routine on every bot. iPhone pause is enough.
Do not retry. Do not start a manual run to see if it works now.
Write the 50-cap owner list: bot name, routine count (max 50),
tightest clock, last output you actually opened.
At a desktop, delete the bot whose tightest clock
is five minutes. Copy the wording first if you still want it.
Deleting removes its routines. It does not sign out Gmail.
Leave every other bot paused. Do not resume from the phone.
Do not hunt a per-bot spend-cap slider. There is not one.
Do not invent a dollar figure for the weekly pool.

OWNER LIST (one row per card)
Bot:
Routines on this card (max 50):
Tightest clock:
Last output I opened:
Decision this hour: paused / deleted poller / still paused

SURVIVOR CLAUSE (paste into every remaining charter)
If a run fails because usage is exhausted, quota is exceeded,
or the product asks you to retry later, stop. Do not retry.
Do not start an alternate route to the same job.
Do not treat this as a network fault.
Write one line: pool empty, paused, waiting for you.
Never initiate spend, a credit top-up, a paid upgrade, or a
purchase to keep working.
Never send, pay, or publish while the weekly pool is empty.
Never set a standing clock tighter than twice a weekday
for inbox work, and never every five minutes.

You decide overflow, in the product, as a person. The bot reports that the pool is empty and it waits. Put the same clause on bots that were not the poller. The briefing bot is the one you will retry. Inbox Triage still never sends when you rebuild it. Standup Scribe still never posts to a shared channel.

Diagnose a Tuesday silence from the roster map, not from a bill

The invoice, if overflow billed, will not arrive in this hour. The symptoms will. Match each one to a single move. Do not debug connectors until every clock is paused.

If even one bot still completes a tiny manual run, you may not be in an empty pool. If every bot fails with the same quota language, you are, and connector repair can wait.

Prove the bleed has stopped with one check that can fail

Waiting twenty minutes is not a check. The weekly pool is weekly. A successful briefing is also not the check, because that success may be overflow you just bought.

Write four facts: the time you paused every routine, the owner list with clocks, the time you deleted the five-minute inbox bot, and whether the product said overflow was available. Then look. Do not run a test fire. The check is absence of the clock.

Open the sidebar. The inbox poller card is gone. Briefing and standup still exist, still paused. No remaining card has a five-minute schedule. The pause list in your notes matches the sidebar. If any row fails, the bleed is not stopped.

Failed rows look like this. The inbox bot is still listed. A copy named inbox-2 is on five minutes. You resumed briefing from the phone. You deleted standup by mistake. You never wrote the owner list, so temp might still hold a tight clock. Fix the failed row. Do not add a manual run to make sure. A manual run is spend.

After a passed check, stay paused until you have a written weekday clock and a desk. Resume nothing from the phone. Rebuild inbox later as a new named owner, twice a weekday. That create flow is the schedule how-to, not this page.

Answer the case for pausing forever and never deleting the poller

The strongest case against deleting the five-minute inbox bot is simple. Pause already stopped the clock. The charter took a week. Labels were useful on Monday. Deleting throws away the card, the chat, and the 20 run records. Resume on Friday, coarsen to twice a day, keep the work. Hide it until then if the sidebar bothers you. Overflow, if the plan allows it, is just paying for speed.

That argument wins for a short, named window: a launch morning, you at the desk, clock back to twice a day before you leave the chair. Duplicate first if you want the wording, then coarsen the copy in the same sitting, then delete or disable the five-minute original before you stand up. Event triggers, where the product offers them, fire when mail arrives rather than 288 times on a dead night. Confirm trigger types in the app.

It loses as a standing plan. Resume is one tap. You will tap it Wednesday when a lead feels urgent. Hide does not pause. Coarsen-later on the same card is how October still has a five-minute inbox. The weekly pool is unpublished, so you cannot prove hundreds of weekday fires will fit. You find out when Chief of Staff Briefing misses Tuesday. Deleting the poller makes five minutes a create action. Create needs a desk, a named owner, and a schedule you type. That friction is the point.

Leave trial credit, overnight incidents, and attended sweeps off this script

This playbook assumes a paid eligible plan, a weekly pool, and unattended routines. Three situations sit outside it.

Trial credit. The one-time trial is limited usage, not a weekly allowance with a published overflow slope. Do not invent a dollar figure for the credit. The trial page is the right place. If you see quota language on a trial, stop and read the terms on the screen. Pause still applies. Delete of a five-minute poller still applies. Overflow assumptions do not.

Overnight incidents. Pause every bot that is not the incident bot, then watch the remaining one yourself. A warm fleet is how the pool dies at 03:00. Daily clocks on Churn Watch and Lead Scout still fire. Pause them.

Attended research. A long public-page sweep you are watching, once, is not a routine. If it hits quota language, stop. Do not convert it into a scheduled job so it can finish overnight. Teach-by-demonstration records up to ten minutes, no microphone audio, browser workflows only, unavailable on iPhone, and it produces a draft skill. That is work. Work spends the pool.

Platforms do not change the pool. Switching to iPhone does not refill the week. Whether Grok Bot is worth another week waits until the bleed is stopped: is Grok Bot worth it. Eligibility can wait too: the Cursor account page.

Make deleting the poller the hard stop you own this week

In the hour you stop grok bot overspending, you are the one who needs a boundary. The product will not cap you. Overflow may keep serving work. Resume will look like diligence. The line is pause every routine, list the 50-cap owners, delete the five-minute inbox bot, and refuse retries.

You never resume a five-minute inbox routine in the same week you hit the empty pool. You never leave that card paused for a future self who will tap resume. You never let the bot initiate spend to keep working. You never send, pay, or publish to catch up while the pool is empty. You never delete briefing or standup as a substitute for deleting the poller. You never invent a dollar figure so the week feels budgeted.

The bot's line is the survivor clause: stop, report, wait. Inbox Triage still never sends. Mail Cleanup Assistant still never permanently deletes mail. When the weekly pool returns, resume from the pause list, coarsest clock first, at a desk. Create a new inbox owner only after the old five-minute card is gone.

Keep reading: Grok Bot Quota Exceeded: What That Message Actually Means, Grok Bot Spend Cap: Set the Monthly Limit, Then Budget in the Charter, How to Schedule a Grok Bot Routine That Does Not Fail Silently.

Frequently Asked Questions

How do I stop grok bot overspending in the next hour?

Pause every routine on every bot before you debug the one that looks broken. iPhone pause is enough. Do not retry, and do not start a manual run to see if the pool came back. Write the 50-cap owner list: each bot card, how many routines it holds, and the tightest clock on it. At a desktop, delete the bot whose tightest clock is five minutes. Leave the others paused. There is no Bot-specific spend cap to raise. Retry is how you step onto on-demand without meaning to.

Should I delete the five-minute inbox bot or only pause it?

Pause first, then delete that one card at a desk. Pause stops the next fire. Resume brings five minutes back in one tap, including from iPhone, which cannot delete or edit. Deleting the poller removes its routines so the standing loop cannot return by muscle memory. Duplicate first if you still want the wording, then disable or delete the live five-minute owner. Do not delete briefing or standup to save the pool. Deleting does not sign out Gmail or clear shared-computer files. It is a spend stop, not a teardown.

Why can three bots empty the weekly pool by Tuesday?

Because they share one unpublished weekly allowance, and one of them can fire hundreds of times before Tuesday morning if its clock is five minutes. A daily briefing and a daily standup might run once each on Monday. An inbox bot on five minutes runs twelve times an hour, including overnight. The product has no Bot-specific spend cap and no audit view outside Enterprise that names the poller, so the tell is a quiet Tuesday briefing, not an inbox warning. The count is the problem. No article can honestly print a dollar size for that pool.

Does pausing or deleting refill the weekly pool this week?

No. Pause stops the next attempt. Deleting a five-minute owner stops that clock from being resumed. Neither refills the included weekly allowance, and a plan change is not documented as a mid-week refill. If overflow is allowed on the plan you hold, further work bills on demand from model and token cost. If overflow is not allowed, work waits until the pool returns. Confirm that in the product. One attended manual run is a decision. Repeated retry is how a paused roster starts spending again.