[BOT ISSUE] Mistral Audio APIs (`audio.speech` and `audio.transcriptions`) not instrumented

[braintrust-bot]

Summary

The @mistralai/mistralai TypeScript SDK provides stable audio APIs for text-to-speech (client.audio.speech.complete()) and speech-to-text (client.audio.transcriptions.complete()). These are model execution endpoints using dedicated Mistral audio models (e.g., voxtral-mini-latest). This repo has zero instrumentation for any Mistral audio surface — no channels, no plugin handler, no wrapper proxy, and no auto-instrumentation config.

What instrumentation is missing

• Wrapper (js/src/wrappers/mistral.ts): Only proxies chat, embeddings, fim, and agents properties. No proxy for audio.
• Auto-instrumentation config (js/src/auto-instrumentations/configs/mistral.ts): No config for any audio.* methods.
• Channels (js/src/instrumentation/plugins/mistral-channels.ts): No channel definitions for audio APIs.
• Plugin (js/src/instrumentation/plugins/mistral-plugin.ts): No handler for audio calls.
• Vendor types (js/src/vendor-sdk-types/mistral.ts): No audio property on MistralClient.

A grep for audio or speech or transcription across js/src/wrappers/mistral.ts and js/src/instrumentation/plugins/mistral-plugin.ts returns zero matches.

Why this is a concrete gap

Both APIs are generative model execution endpoints:

SDK Method	Description	Model
client.audio.speech.complete()	Text-to-speech synthesis — generates audio from text input	Dedicated TTS model
client.audio.transcriptions.complete()	Speech-to-text — generates text transcript from audio input	voxtral-mini-latest

The transcription API supports features like speaker diarization, word-level timestamps, language specification, and context biasing. These are not CRUD endpoints — they invoke models and produce generated output.

Braintrust docs status

not_found — The Braintrust Mistral integration page at https://www.braintrust.dev/docs/integrations/ai-providers/mistral documents chat completions only. Audio APIs are not mentioned.

Upstream references

• Mistral Speech-to-Text docs: https://docs.mistral.ai/capabilities/audio/speech_to_text
• Mistral Text-to-Speech docs: https://docs.mistral.ai/capabilities/audio/text_to_speech
• Mistral Audio Transcription API endpoint: https://docs.mistral.ai/api/endpoint/audio/transcriptions
• Mistral TypeScript SDK: https://github.com/mistralai/client-ts — client.audio.speech and client.audio.transcriptions resources documented in README

Precedent in this repo

• OpenAI audio APIs (audio.transcriptions.create() and audio.speech.create()) are tracked as a gap in issue OpenAI audio.transcriptions.create() and audio.speech.create() not instrumented (wrapper and auto-instrumentation) #1629.
• This repo instruments 8 other Mistral SDK methods across 4 API categories (chat, embeddings, FIM, agents).

Local files inspected

• js/src/wrappers/mistral.ts
• js/src/auto-instrumentations/configs/mistral.ts
• js/src/instrumentation/plugins/mistral-channels.ts
• js/src/instrumentation/plugins/mistral-plugin.ts
• js/src/vendor-sdk-types/mistral.ts
• e2e/scenarios/mistral-instrumentation/