Constant error after latest update

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

Request ID: 841107d5-3a82-4df6-924e-b640ee46f4c5
The agent stopped because of an internal error. Try again.
RetriableError: The agent stopped because of an internal error. Try again.
at __b (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:9085:128032)
at fGo._executeTurn (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:9085:150068)
at async fGo.executeTurn (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:9085:138335)
at async Sme.runAgentLoop (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:9080:17083)
at async PRa.streamFromAgentBackend (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:19013:24590)
at async PRa.getAgentStreamResponse (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:19013:49924)
at async $Ie._submitChatMaybeAbortCurrent (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:9063:5198)
at async $Ie.submitChatMaybeAbortCurrent (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:9058:90558)

Steps to Reproduce

After every prompt I get “The agent stopped because of an internal error. Try again.”

Operating System

Windows 10/11

Version Information

Version: 3.27.22 (system setup)
VS Code Extension API: 1.128.0
Commit: b1aa50cb6db84fd3e8c808259cb172ea73425c10
Date: 2026-10-27T00:33:02.045Z
Layout: IDE
Build Type: Stable
Release Track: Default
Electron: 42.11.16
Chromium: 148.0.7778.280
Node.js: 24.19.0
V8: 14.8.178.38-electron.0
xterm.js: 6.1.0-beta.291
OS: Windows_NT x64 10.0.26100

Does this stop you from using Cursor

Sometimes - I can sometimes use Cursor

Hi @defaultroot,

This doesn’t look like it’s caused by the update. Your requests reach our servers and get answered normally, but Cursor then fails to save the chat to its local database on your machine, which reports it as damaged (“database disk image is malformed”), and the agent stops with that generic error. Cursor doesn’t repair that database on its own yet, so restarting usually won’t clear it.

First try a brand-new chat. If that fails the same way, resetting the local database should help:

  1. Quit Cursor completely (check Task Manager that no Cursor processes are left).
  2. Open %APPDATA%\Cursor\User\globalStorage in File Explorer.
  3. Copy state.vscdb (and state.vscdb-wal / state.vscdb-shm if present) to another folder as a backup.
  4. Rename the originals, for example to state.vscdb.old.
  5. Start Cursor, sign in again if asked, and send a prompt in a new chat.

Your local chat history won’t show after this, but it stays in the backup copy. Let me know whether prompts go through after the reset.

I can’t tell the cause from the stack trace alone, but when every prompt fails right after an update, these are the things I’d check first:

  1. Look at https://status.x.ai to rule out an outage
  2. Go to Cursor Settings > Network > Run Diagnostics and see if anything fails
  3. If you’re behind a VPN or corporate proxy (Zscaler etc.), set HTTP Compatibility Mode to HTTP/1.1 and fully restart Cursor
  4. Start a fresh chat and try a different model to see if it’s tied to one model or chat

If it still fails, post what Run Diagnostics shows. The request ID you already shared is exactly what the team needs.

more deets: Network, proxy, and remote connections | Cursor Docs
bug report tips: Reporting a bug | Cursor Docs

Is not grok

That worked. But, is not the first time that happens

Fair, status.x.ai is just where status.cursor.com redirects now, so it covers Cursor too. Sounds like this one was the local chat database, not an outage, so glad Colin’s reset fixed it.