Repeated Grok Bot computer freezes

Hi Cursor team,

First, I really like what Grok Bot makes possible. I’ve built an assistant on it that runs scheduled, time-sensitive tasks on its computer throughout the day, with careful guardrails and my sign-off on any changes. When it runs smoothly, it saves me hours every week, and that’s exactly why I’m a paying customer.

Unfortunately, the last week has been rough on reliability, and I haven’t been able to get a human to look at it:

Repeated computer freezes. The bot’s computer keeps stalling and then jumping forward in time. On Nov 4 alone it froze 13 times in one evening, about 3.5 hours of downtime, including stretches of 53 and 50 minutes. There were more freezes earlier that day and earlier in the week.

Missed and late tasks. While the computer is frozen, scheduled tasks simply don’t run. That evening, eight time-sensitive tasks were missed outright and three more ran late.

Usage drain. My weekly usage jumped sharply in a single day, much of it spent restarting, rerunning interrupted work, and troubleshooting after the freezes, even after I cut my schedules way down to save usage.

No human response yet. Ticket T-I63839 has been open for a full week. I’ve sent detailed, dated incident logs but have only received automated replies.

Could someone from the team please take a look at ticket T-I63839 and reply directly? I’m happy to share the full logs with timestamps for any of these outages.

I want to keep building on Grok Bot, and I appreciate the team’s work on it. Thank you for taking a look.

Thanks,

Anthony

Hi there!

We detected that this may be a bug report, so we’ve moved your post to the Bug Reports category.

To help us investigate and fix this faster, could you edit your original post to include the details from the template below?

Bug Report Template - Click to expand

Where does the bug appear (feature/product)?

  • Editor, Tab & Chat (autocomplete, Composer, in-editor agent)
  • Terminal & commands
  • Models, pricing & API keys (availability, Auto/Max, BYOK/Bedrock)
  • MCP & tools
  • Cloud Agents & Automations (cursor.com/agents, scheduled/event)
  • BugBot & Code Review
  • Cursor CLI
  • Cursor Mobile
  • Remote (SSH / Dev Containers / WSL)
  • Account, billing & login
  • Something else…

Describe the Bug
A clear and concise description of what the bug is.


Steps to Reproduce
How can you reproduce this bug? We have a much better chance at fixing issues if we can reproduce them!

  1. …
  2. …
  3. …

Expected Behavior
What is meant to happen here that isn’t working correctly?


Screenshots / Screen Recordings
If applicable, attach images or videos (.jpg, .png, .gif, .mp4, .mov)


Operating System

  • Windows 10/11
  • MacOS
  • Linux

Version Information

  • For Cursor IDE: Menu → About Cursor → Copy
  • For Cursor CLI: Run agent about in your terminal
IDE:
Version: 2.xx.x
VSCode Version: 1.105.1
Commit: ......

CLI:
CLI Version 2026.01.17-d239e66

For AI issues: which model did you use?
Model name (e.g., Sonnet 4, Tab…)


For AI issues: add Request ID with privacy disabled
Request ID: f9a7046a-279b-47e5-ab48-6e8dc12daba1
For Background Agent issues, also post the ID: bc-…


Additional Information
Add any other context about the problem here.


Does this stop you from using Cursor?

  • Yes - Cursor is unusable
  • Sometimes - I can sometimes use Cursor
  • No - Cursor works, but with this issue

The more details you provide, the easier it is for us to reproduce and fix the issue. Thanks!

Not a bug report

Hi @scrolling2024, the freezes you’re seeing are the computer going to sleep, not crashing. When no Bot is actively working on it and no app is using it, it pauses a few minutes later and picks up where it left off on the next wake. Anything running inside the computer (scripts, its own schedulers, cron) doesn’t count as activity, so it freezes too. That’s why the clock jumps and why tasks scheduled inside it get missed.

Your Nov 4 timeline matches this, including the 53 and 50 minute stretches. This is a known issue the team is aware of, and keeping the computer awake while background work runs is something they’re looking at. There’s no way to keep it on 24/7 today.

In the meantime:

  1. Move time-sensitive jobs into scheduled Bot routines rather than schedulers inside the computer. A routine runs on time and wakes the computer when it needs it.
  2. For anything that has to stay inside the computer, have it check on wake how much time has passed and catch up, so a pause doesn’t lose a run.

Hi Colin, thanks for the fast and clear reply. It’s the first real explanation I’ve gotten, and I appreciate it.

One thing I want to flag: the bot wasn’t idle during these freezes. I already run the time-sensitive work the way you suggest, as scheduled Bot routines rather than only schedulers inside the computer. On Nov 4 evening, a routine was scheduled to fire every 15 minutes (:10, :25, :40 and :55), plus a passive check every 30 minutes, and I was actively chatting with the bot through much of it. Even so:

Several freezes started or ran right across scheduled routine times. Examples: 16:24–16:36 spanned the 16:25 run, 18:43–19:02 spanned 18:55, 19:05–19:12 spanned 19:10, 19:15–19:27 spanned 19:25, and 20:37–20:42 spanned 20:40 (all JST).

The routines themselves were delayed by the freezes instead of waking the computer on time. The 17:55 run didn’t start until 18:04, the 18:10 run started at 18:17, and the 19:55 run started at 19:58.

With activity at least every 15 minutes, a “pause a few minutes after idle” can’t explain a 53-minute freeze (17:11–18:04) or a 50-minute one (20:47–21:36).

So it looks like the computer is pausing even while routines and chats are active, and the scheduled routines aren’t reliably waking it. I’m also building in catch-up on wake as you suggested, but right now missed time can’t be recovered for time-sensitive work.

Could the team check whether routine wakes are being delayed on my account, and let me know if there’s a setting or workaround? I’m happy to share the full logs. Thanks again for looking into this.

Hi @scrolling2024,

I may have oversimplified things a little. Every scheduled routine starts at a random point up to 10 minutes after its set time, which is why 17:55 started at 18:04 and 18:10 at 18:17. A routine does wake the computer, but only once it starts and uses it, and the computer can sleep again a few minutes after the run ends. Chat replies work the same way: a reply that doesn’t use the computer doesn’t wake it.

@Colin, thanks for explaining that internal scripts don’t keep the computer awake and that scheduled routines can start up to 10 minutes late. We have implemented bot-scheduled wake routines and catch-up handling, but dependable operation is still unresolved.

Two examples from today, November 6, all times JST:

A chart read scheduled for 09:55 started at 10:03:20, after its 10:02 deadline.

A keep-awake hold intended to begin at 10:28 did not start until 10:32:36.

These delays fit your explanation, but they leave a practical question: how should a customer reliably perform time-sensitive work when the wake-up itself may arrive after the deadline? We are considering starting preparation earlier and refreshing the decision closer to entry, but have not tested that change yet. Is keeping that same routine active until its final check a supported approach, and what actually counts as activity during that interval?

We found and corrected mistakes in our own configuration, and I’m not attributing every missed trade or legitimate no-trade decision to Cursor. Nevertheless, too much of this project is now spent troubleshooting basic execution instead of doing productive work—even after upgrading to Super.

I have been seeking substantive support answers and a credit resolution for over a week. Ticket T-I63839 still has no clear engineering owner, resolution timeline, or compensation decision.

Please help me escalate this to someone who can provide a concrete, supported operating approach and a firm update on the credit request. I appreciate your help.