Continuing without info from webpage

[Danack (Dan Ackroyd)]

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

Cursor appears to continue even when it has failed to read info.

Cursor didn’t know how to read the title of a video through an embedded YouTube player, and was thinking about writing some server side code with a new API endpoint. I stopped, went to ChatGPT and asked it, which gave what looks like valid code, so I grabbed the link to the conversation.

I then asked Cursor to read the ChatGPT conversation and use the code from it. Cursor failed to read it, and then merrily tried to continue implementing the backend API.

Steps to Reproduce

Probably ask it to read a chatgpt conversation

Expected Behavior

When explicitly told to read a webpage to gain some knowledge, Cursor should say it has failed to read a webpage rather than continuing without the knowledge.

Screenshots / Screen Recordings

Operating System

MacOS

Version Information

Version: 2.6.11
VSCode Version: 1.105.1
Commit: 8c95649f251a168cc4bb34c89531fae7db4bd990
Date: 2026-03-03T18:57:48.001Z
Build Type: Stable
Release Track: Default
Electron: 39.6.0
Chromium: 142.0.7444.265
Node.js: 22.22.0
V8: 14.2.231.22-electron.0
OS: Darwin arm64 25.2.0

For AI issues: which model did you use?

Auto

For AI issues: add Request ID with privacy disabled

473fc11b-7b12-4c28-9d7a-304b9c47b719
b01d7205-301d-4af9-9849-9f823dc812a5

Does this stop you from using Cursor

No - Cursor works, but with this issue

[deanrie]

Hey, thanks for the report. Two things are going on here:

1. ChatGPT pages are client-rendered SPAs: the fetch tool can only grab the initial HTML, which for ChatGPT is basically an empty shell. The actual conversation content loads via JavaScript, so the fetch returns almost nothing useful. This is a known limitation for JS-heavy pages.

2. The real bug: you’re right that the agent should’ve told you it failed to get meaningful content instead of silently pressing on.

As a workaround for now, instead of linking to a ChatGPT conversation, you can copy-paste the relevant code or response directly into the chat. That way the agent gets the full context right away.

Let me know if anything else comes up.