clarify model origin machine - #224
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 3119e4f101
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| key: "config/ai/transport_target", | ||
| label: "Origin machine", | ||
| description: "Where custom provider HTTP requests start: the GSV Worker or a machine with network fetch support.", | ||
| description: "Where model provider HTTP requests start: the GSV Worker or one of your machines. Use a machine when the provider blocks Cloudflare Worker IPs.", |
There was a problem hiding this comment.
This description tells users that setting an origin machine changes where model-provider HTTP requests start, but the kernel only builds a routed fetch when the provider is openai-codex or shouldUseCustomProvider(...) is true; built-in providers like anthropic, openai, or openrouter still run through the default generation path even if transport_target is set, as covered by the existing does not build a routed fetch for non-custom text generation targets test. In the provider-blocks-Cloudflare case described here, a user can follow this UI guidance for a built-in provider and still have traffic originate from the Worker.
Useful? React with 👍 / 👎.
1 participant
No description provided.