Cursor completely unusable with "Taking longer than expected" error

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

Any request to any model is resulting in a “Taking longer than expected”. I have restarted my machine, downgraded HTTP compatibility mode to HTTP/1.1 and HTTP/1.0, tried in a new blank folder and run the network diagnostics (that report some errors). Have also changed network connections but to no avail. Cursor is completely unusable right now.

Steps to Reproduce

Submit a chat request to plan or to make code changes.

Screenshots / Screen Recordings

Operating System

MacOS

Version Information

Version: 2.5.26 (Universal)
VSCode Version: 1.105.1
Commit: 7d96c2a03bb088ad367615e9da1a3fe20fbbc6a0
Date: 2026-02-26T04:57:56.825Z
Build Type: Stable
Release Track: Default
Electron: 39.4.0
Chromium: 142.0.7444.265
Node.js: 22.22.0
V8: 14.2.231.22-electron.0
OS: Darwin arm64 25.3.0

For AI issues: which model did you use?

Opus 4.6, Composer 1.5, GPT-5.3 Codex

Does this stop you from using Cursor

Yes - Cursor is unusable

Hey, thanks for the report. From the diagnostics, I can see DNS, downloads, CDN, and most endpoints are OK, but Authentication fails with a “Failed to fetch” error. That’s likely the root cause. If authentication doesn’t work, all AI requests will time out.

A couple questions:

  1. Are you behind a corporate proxy, VPN, or firewall? That’s the most common reason why Authentication fails while other endpoints still work.
  2. Can you try a different network, like a mobile hotspot, to rule out filtering on your corporate network?

I also see you’re on version 2.5.26. The latest is 2.6.11. It’s worth updating since we fixed a lot of bugs.

A few more things to try:

  • Sign out and sign back in: Cursor Settings > sign out, then sign in again.
  • Grab the Request ID from a failed chat request (top-right of the chat panel > menu > Copy Request ID) and send it here. We need it to trace what’s happening on the backend.

If you’re on a corporate network, your IT team may need to allowlist prod.authentication.cursor.sh and *.cursor.sh. Docs are here: Network Configuration

Let me know what you find.

1 Like

I tried signing out and signing in again with the current version but thats not working either. Now i am unable to sign in the app when I am authenticated on the web and can see my account details. In the app, it just remains stuck to Sign in.

The update within the app is not working either. If i try to download the upgraded version from the website, it takes me to this page: https://api2.cursor.sh/updates/download/golden/darwin-arm64/cursor/2.6 which is down currently as well. I have tried multiple internet connections including my personal non firewall/vpn one.

The app was working like a breeze last night. I had two instances open and was working heavily on 2 apps at the same time. What is going on now ?