Mar 19
Feature Request: Native Integration of Mistral AI's Voxtral Models
Dear Superwhisper Team,I would like to formally request the integration of Mistral AI's Voxtral model family as a supported transcription provider within Superwhisper, across all platforms — macOS, Windows, and iOS.Mistral AI recently released Voxtral Transcribe 2, currently their most advanced speech-to-text offering, comprising two models designed for distinct use cases:• Voxtral Mini Transcribe V2 — a batch transcription model that achieves approximately 4% word error rate on the FLEURS benchmark, outperforming GPT-4o mini Transcribe, Gemini 2.5 Flash, AssemblyAI Universal, and Deepgram Nova on accuracy, while processing audio roughly 3× faster than ElevenLabs Scribe v2 at approximately one-fifth of the cost ($0.003/min via API).• Voxtral Realtime — a streaming model with configurable latency down to sub-200ms, released as open weights under the Apache 2.0 license, making it particularly attractive for privacy-conscious deployments and real-time dictation workflows.Both models offer native multilingual support across 13 languages, including Portuguese, French, German, Spanish, and Arabic, among others — a meaningful advantage for your international user base.For Superwhisper users, this integration would deliver a compelling combination of best-in-class transcription accuracy, significantly lower API costs compared to existing cloud providers, and — in the case of Voxtral Realtime — the possibility of self-hosted or on-device deployment for users with privacy requirements.Given that Superwhisper already supports Bring Your Own Key (BYOK) for OpenAI-compatible providers, I would also encourage consideration of extending this mechanism to accommodate Mistral's API format (api.mistral.ai), which would unlock Voxtral access for Pro users immediately, without requiring deep platform-level integration.I believe this addition would meaningfully expand Superwhisper's appeal to users seeking high-quality, affordable, and open transcription — and would position Superwhisper as one of the few voice-to-text tools to offer access to what is arguably the best open-weight speech model available today.Thank you for your continued work on this product. I hope this suggestion receives consideration.Best regards
Pending