ChatGPT
ChatGPT copied to clipboard
lencx
[Bug] Stuck at cloudflare check.
Version
0.8.0
Bug description
Using latest Appimage on fedora 37. I can use the website through a browser fine. Firefox and chrome both work. But this app gets stuck at the cloudflare check. I tried changing the user agent to match the one I use in chrome, but that did not fix anything.
OS
Fedora 37
Environment
I would try building from source but I notice that VS Code is a requirement. Can I build it with VSCodium or Code OSS?
It does not require a specific code editor, rust and node.js environments are required.
Cloudflare is a network problem, this problem is very frequent in China, many people have feedback on this problem, it comes from the openai website security policy, and the application does not have much to do with.
I see, I'll try to build it then. Just the build instructions on the readme say to pre-install vscode and some extension: https://github.com/lencx/ChatGPT#how-do-i-build-it
I'm in UK. And the site works fine in other browsers.
tauri is calling the system's built-in browser, and I can't be sure what detection mechanism openai is.
Do you mean the default browser? So that if I change the default browser setting on my system it will change for this?
I'm not sure, you can try (I don't think it has anything to do with the computer's settings).
tauri is calling the system's built-in browser, and I can't be sure what detection mechanism openai is.
I'm on kde, so QTWebkit? There must be a reason they are treating you different to other clients. Let me know if there's something I can do to help you diagnose it.
Re:
Cloudflare is a network problem, this problem is very frequent in China, many people have feedback on this problem,
Interesting. AFAIK, also they are bound to apply AI network and even end-point filtering (mandated by the US government, for starters), as per OpenAI's ToUs, see my comment there or read them directly there.
I am also stuck on the Cloudflare check. @irthomasthomas did you manage to fix this problem?
FYI, it has been working for me, almost always, for the past +2 weeks, better than in the "bare" browser, even despite intermittent:
Due to high demand on our systems, previous conversations are temporarily unavailable
We're experiencing exceptionally high demand. Please hang tight as we work on scaling our systems.
EU here: I half suspect they implemented regional throttling, partly as per their ToUs, see above.
In short, in my EU country now @lencx 's tool is more reliable than the non-API direct access.
As an (not topical) aside : the newest GPTChat (directly at backend) feature "stop generating" is a godsend.
It saves generation tokens and patience: win-win situation.