Agent review on commit fails

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

When I commit, I can know that the agent review process starts, but it fails and shows the message: “Failed to run review: insufficient funds (request ID: 67d362da-8645-45f6-a224-7c44a6387f79)”.

Steps to Reproduce

  1. Commit my code
  2. Wait a sec
  3. Failed to run review tip appears

P.S. This works great the last several days, but cannot review my code today.

Expected Behavior

Work correctly

Screenshots / Screen Recordings

Operating System

Windows 10/11

Version Information

Version: 3.9.16 (system setup)
VS Code Extension API: 1.105.1
Commit: 042b3c1a4c53f2c3808067f519fbfc67b72cad80
Date: 2026-06-27T06:41:01.941Z
Layout: editor
Build Type: Stable
Release Track: Default
Electron: 40.10.3
Chromium: 144.0.7559.236
Node.js: 24.15.0
V8: 14.4.258.32-electron.0
xterm.js: 6.1.0-beta.256
OS: Windows_NT x64 10.0.26200

Does this stop you from using Cursor

No - Cursor works, but with this issue

Hey, this isn’t a bug, it’s expected behavior. Agent Review runs on server-selected models, and it uses usage the same way a normal agent request does. On the free plan, once your included quota runs out, review requests hit the limit and fail with insufficient funds, which is exactly what you’re seeing in the screenshot.

The fact that it worked for a few days and then stopped also matches running out of the free quota.

What you can do:

  • Upgrade to Pro, or enable usage-based on-demand spending so review keeps working after your included quota is used up.
  • Switch review to Quick mode in Cursor Settings. It’s cheaper per run, but it still uses quota.
  • The free quota resets at the start of the next billing cycle.

There’s no auto model option for Agent Review. It’s a separate pipeline from the normal agent, and you can’t manually switch the model there. More details: Agent Review | Cursor Docs

Hey Dean, I’m not on a free plan, I am on the ultra plan. But this issue is happening in a different account. My original account ran out of quota, and I switched to another account this morning. After switching, I started seeing that error.


And my friend says the API quota has ran out few days ago, but showing number is 0 percent. Is agent review having anything to do with API quota?

Thanks for the details and the screenshot. Let me break it down.

About the API quota, yeah, it’s related. Agent Review is a separate pipeline. It runs on server-selected premium models. So it uses your premium or API usage, not the Auto + Composer bar you see at 40%. So your friend is right overall, Review is hitting the premium or API limit.

The key part in your screenshot is On-Demand Spending: Disabled. Once your included premium or API allowance is used up, and on-demand is off, Review has nothing to charge to and it fails with insufficient funds. The dashboard percentages can also lag, so API 0% does not always reflect the real-time state. If you want Review to keep working after you hit the included limit, turn on usage-based on-demand spending.

About switching accounts. The error comes from the account the IDE is signed into, not the one you have open in the browser. After switching accounts, fully sign out in Cursor Settings > sign out, sign back in, and restart Cursor. Otherwise the editor may still be using the old account that ran out of premium usage, and then the Ultra dashboard you’re looking at is not relevant.

One more thing about yesterday. We had an incident. That issue showed “We’re experiencing high demand right now. Please try again in a few moments.” That’s Auto load and a different system. Your error is insufficient funds, which points to usage and spend limits, not that incident.

If after re-login and enabling on-demand on Ultra you still get insufficient funds while usage is clearly low, email [email protected] and include the Request ID from the failed Review request from that account, and they can check your specific case.

That’s terrifying to know agent review is consuming API quota. I’d rather to run bugbot subagent locally with Grok 4.5. :joy:

Thanks a lot for pointing this out, really appreciate your help!