Fix Z.ai and MiniMax provider inference#146
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR stabilizes the new Z.ai and MiniMax provider paths.
It adds provider-specific model catalogs and safe model selection so saved OpenAI model ids do not leak into Z.ai or MiniMax requests. Z.ai model listing now falls back to the Coding Plan model list when the endpoint is unavailable or unhelpful.
MiniMax requests now avoid browser-side OpenAI SDK calls that trigger CORS failures from Stainless headers. Model discovery, primary chat completions, summarization, and context analysis go through the Electron HTTP bridge instead.
The MiniMax chat path also filters reasoning output before it reaches chat or TTS, avoids unsupported
systemrole messages in tool-call continuations, and fixes the audio-state race where synthetic one-shot responses could flipSPEAKINGback toIDLEbefore playback actually started.Impact
Users can select valid Z.ai and MiniMax models in settings, use MiniMax without CORS failures, continue tool-call conversations, and get cleaner spoken/chat responses without reasoning text leaking into the UI.