Grok Bot 0430 Mac desktop session broken iOS works

Support Bug Reports
,
1

Account email: 【写成 名字 at 域名 dot com】
User id: google-oauth2 user_01KN64F1EFSPR7EZFTJSSVGTVV

Please reattach this Mac desktop session to my existing Agent Computer.
Do not rebuild. iOS still works and the bots are there.

Mac Grok Bot 0430 cannot attach. Retry does nothing.
In-app Send Feedback also fails on this Mac.
I did not tap Restore.

Cloud computer is not down. Phone works.
The 0430 firewall copy is wrong. This Mac can reach the us9 cursorvm host and api2. TLS ok. GET returns 404 which staff said is normal.

Box host (replace dots yourself when you look it up):
1b385416305d464f8fb3-pod-wxpkq5gbj5a4rfuogrcciek5my-1340 us9 cursorvm com

Logs from 0300: cursor-profile ConnectError, base_url unknown, ControlPortCallError.
Local profile has no gateway descriptor.

Please reattach the Mac gateway binding. Keep iOS bots.

2

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!

5

Account email: 【填写你的邮箱】
Cursor / Grok user id: google-oauth2|user_01KN64F1EFSPR7EZFTJSSVGTVV

Please reattach this Mac desktop session to my existing Agent Computer.
Do NOT rebuild/reset unless necessary. iOS still works and the bots are there.

What I see

  • iOS Grok Bot: same account, works.
  • Mac Grok Bot 0.43.0 (upgraded from 0.30.0): cannot attach to the existing cloud computer.
  • Current UI: cannot connect to your computer. Retry plus Restore Computer. Retry does nothing.
  • In-app Send Feedback also fails: cannot submit this report.
  • I did not tap Restore or Reset.

This is not the cloud computer being down. If it were, iOS would fail too.
This is not firewall / VPN / Zscaler blocking star.cursorvm.com. That 0.43.0 error copy is wrong on this Mac.

The Mac app gateway/session path is broken. That is why both connect-to-computer and submit-feedback fail on desktop only.

Environment

  • Grok Bot desktop 0.43.0 stable, darwin arm64
  • Previously 0.30.0 stuck on black screen plus white loading circle after an update on Aug 30
  • macOS 26.4.1, Apple M4
  • Cursor IDE on this same Mac works
  • Timezone Asia/Shanghai UTC+8

Investigation

  1. 0.30.0 logs on Aug 30
  • desktop startup ready
  • edge_failed area cursor-profile, ConnectError.Unknown
  • base_url_kind unknown
  • box_reachability ControlPortCallError hundreds of times
  • box_secrets.push host_unreachable
  1. Forum match
  • Same symptom: iOS OK, Mac loading/black after update
  • Staff on topic 169940: cloud computer is healthy; Mac desktop cannot open a session
  • Staff: rebuild will not help if the box is fine
  • No product-wide fix announcement as of Sep 5
  1. After upgrade to 0.43.0 on Sep 5
  • First screen blamed firewall/VPN/Zscaler and star.cursorvm.com
  • Later screen: cannot connect to your computer, bots are still there
  • Local profile has no gateway descriptor
  • In-app feedback submit fails on the same desktop channel
  1. Network checks. The firewall story is false.
    This account box hostname is already visible on this Mac:
    1b385416305d464f8fb3-pod-wxpkq5gbj5a4rfuogrcciek5my-1340.us9.cursorvm.com
  • Resolves, TLS 1.3 OK, cert CN us9.cursorvm.com, SAN star.us9.cursorvm.com
  • GET slash returns 404, same as test.us8.cursorvm.com (staff said 404 means connect worked)
  • api2.cursor.sh returns 200
  • Versioned 0.43.0 dmg on downloads.cursor.com sand path returns 200
  • In-app update index on that CDN returns S3 403 listing denied. That is why check-update fails. I already installed 0.43.0 manually.

So this Mac can reach the us9 cursorvm host and api2. The desktop client still cannot attach a session or submit feedback.

Ask

  1. Reattach this Mac session / gateway binding to the existing Agent Computer above.
  2. Do not rebuild unless you confirm the computer is actually stuck. Keep the iOS bots.
  3. Please check why desktop in-app feedback cannot be submitted on this session.
7

I’m seeing what looks like the same issue on my Mac.

The Grokbot conversation appears to be stuck on the desktop app and will not refresh. The bot itself is still working normally in the cloud, and the same conversation is current and usable in the iOS app.

Desktop symptoms:

  • Existing Grokbot conversation shows “Couldn’t load conversation” / appears stuck
  • Clicking Retry does nothing
  • Messages sent from the desktop may show “failed to send,” but the bot can still receive them server-side
  • The desktop view does not update with the bot’s latest responses
  • Other Grokbot conversations may still load, so this does not appear to be a total connectivity failure

Troubleshooting already tried:

  • Force-quit and restarted the desktop app
  • Signed out and back in
  • Updated Grokbot’s Agent Computer
  • Switched the Mac from normal Wi-Fi to an iPhone hotspot
  • Retried the affected conversation multiple times

None of those changed the behavior.

Meanwhile:

  • Grokbot is still running and doing work in the cloud
  • iOS shows the current conversation correctly
  • Commands sent to the bot are still being received/executed
  • Only the macOS desktop conversation view seems stale / unable to reattach or refresh

This makes it look more like a desktop conversation/session synchronization or attachment issue than an Agent Computer or network problem.

Happy to provide logs or other diagnostics if the team wants anything specific checked.

8

same issue here on 0.43. CoS stopped connecting in macos. iphone is fine. tried rebooting and reinstalling grokbot. no luck

9

Same 0.43.0 Mac bug as this thread. iOS on the same account and same Wi‑Fi is fine. The cloud bots are working.

After I updated the Mac app to 0.43.0 the conversation pane went stale (“showing saved messages / out of date”). Sidebar previews were already newer than the open thread. I fully quit, then signed out and back in. That cleared the cached thread and left “Couldn’t load conversation. Check your connection and try again.” Retry does nothing. It happens on every conversation I open, not just one.

The Mac app is not offline. Sidebar last-message previews update live when I send from the iPhone. The computer/monitor pane works; I can see the bot’s screen. Routines load. I can send from the Mac composer. A test message I sent from Mac appeared in the thread as my own bubble. The bot reply showed immediately in the Mac sidebar and on iOS, but it never appeared in the Mac transcript.

So on desktop, my own outgoing message can render. Old history and incoming bot replies do not. Everything else in the same window is alive.

Already tried: Retry, Cmd+Q and reopen, sign out and sign in, opening a different conversation, same network as the working iPhone. I have not used Reset, Recover, or Update Agent Computer, and I have not created a new bot or reinstalled.

10

Additional diagnostic evidence since my last reply:

I’ve now confirmed this is isolated to ONE specific Grokbot conversation/bot
(Cecil), not a general macOS Grok Bot failure on my account.

What I’m seeing:

  • Cecil’s conversation is still current and fully usable on iOS.
  • Cecil is fully operational in the cloud and continues doing work.
  • Commands sent from the Mac may show “failed to send”, but Cecil still
    receives and executes them server-side.
  • Cecil’s macOS conversation pane remains stale / cannot refresh.
  • Clicking Retry still does nothing.

Most importantly:

  • My OTHER Grokbot conversations/bots on the SAME Mac app load and update
    normally.
  • I can see messages/updates from those other bots on macOS.
  • Cecil can communicate with/update those other bots, and those updates are
    visible on the Mac.
  • So the Mac app itself is not generally offline, and the shared Agent
    Computer is functioning.

Troubleshooting already completed:

  • force quit/relaunch
  • sign out/sign in
  • repeated Retry
  • switched from normal Wi-Fi to iPhone hotspot
  • Updated Agent Computer

None changed Cecil’s desktop conversation.

This now looks much more like a stale/broken macOS attachment or synchronization
binding for this specific conversation/session, rather than a network issue or
Agent Computer failure.

Could the team please inspect/rebind this specific desktop
conversation/session without replacing the bot or destroying its existing
cloud state?

I can provide any account/session/conversation identifiers or logs that would
help isolate it.

11

MAJOR NEW REPRODUCTION / DIAGNOSTIC UPDATE

I attempted to migrate the affected bot (Cecil) to a fresh duplicate called
“Cecil v2” because the original Cecil macOS conversation remained broken.

Cecil v2 initially worked normally on macOS.

It successfully completed a long sequence of interactions, including:

  • reading durable handoff state
  • reconciling repo/orchestration state
  • inspecting its routines
  • reporting current worker/research status
  • inspecting its disabled managed-worker-exit-wake webhook routine

The conversation was completely functional throughout all of this.

Then I used Grokbot’s SECRET INPUT UI to securely enter/save the Webhook URL
for Cecil v2’s disabled managed-worker-exit-wake routine.

The UI showed:

“Cecil v2 exit-wake Webhook URL — Saved securely and kept private.”

Immediately after this secret-input operation, Cecil v2’s macOS conversation
began exhibiting the SAME failure mode as the original Cecil conversation:

  • messages show “Failed to send”
  • transcript/session no longer behaves normally on desktop
  • this occurred immediately after the webhook URL secret-input event

This is especially important because the webhook migration itself had NOT
occurred yet:

  • v2 webhook routine was still disabled
  • webhook key had not yet been supplied
  • shared runtime env had not been changed
  • no worker had been launched
  • no webhook smoke test had occurred

So this does NOT appear to be caused by my webhook code actually firing.

The apparent trigger is the Grokbot/macOS conversation handling around the
webhook SECRET INPUT / webhook credential workflow itself.

This is now a reproducible case:

  1. Create/duplicate healthy bot
  2. Bot conversation works normally on macOS
  3. Perform webhook secret-input operation
  4. macOS conversation immediately develops the same broken send/sync behavior

The original affected Cecil also owns webhook routines, which may now be
relevant to the original failure.

Please investigate the desktop conversation/session handling associated with
Grokbot webhook credentials / secret-input events.

I am preserving both affected bot instances and will not delete them so the
team can inspect their server-side state.

12

Hi @jian_ma @ssllle @Bean @techntrails

Your boxes should all be healthy now. please try reopening Grok Bot and confirm if that’s the case!