Local Audio Transcription with Cohere Transcribe on WebGPU

Melvin Vivas · X post · 2026-04-04 · Open on X

Topics: LLM Fundamentals, LLMOps, Deployment & Monitoring · Level: beginner

Summary

The creator shows Cohere's speech-to-text model running locally in the browser through WebGPU. He links the model card on Hugging Face and a Hugging Face Space where anyone can try it.

Key points

• Cohere released a speech-to-text model: cohere-transcribe-03-2026.
• With WebGPU, it runs entirely in the browser on your own machine, with no server.
• A Hugging Face Space lets you try it with no setup.
• Running transcription in the browser keeps your audio private.

Resources mentioned

• CohereLabs/cohere-transcribe-03-2026 · tool · huggingface.co · free
  Model card on Hugging Face for Cohere's audio transcription (speech-to-text) model.
• Cohere Transcribe WebGPU (Hugging Face Space) · tool · huggingface.co · free
  Browser demo by CohereLabs that runs the transcription model locally using WebGPU.
• Cohere · person · x.com · free · recommended by both Bashiri Smith & Melvin Vivas
  AI company that builds language, embedding and reranking models, and now the Transcribe speech-recognition model.
  Also in: Basic RAG Pipeline in 60 Seconds: From Documents to Grounded Answers (Bashiri Smith on Facebook · notes), Cohere Parse Beats Frontier LLMs at Receipt Parsing (Melvin Vivas on X · notes), Cohere Parse: Pricing vs Parse Bench Score (Melvin Vivas on X · notes), Cohere's North Micro Vision: a small open-source vision model for documents (Melvin Vivas on X · notes) and 1 more
• Hugging Face · website · x.com · free · recommended by both Bashiri Smith & Melvin Vivas
  Platform for hosting and finding ML models, datasets and papers. The quoted post says LocateAnything was trending there.
  Also in: Using an ML agent to train an open-source TTS model on your voice (Melvin Vivas on X · notes), Deploy Open-Source Models with Hugging Face Inference Endpoints (Melvin Vivas on X · notes), Deploying Qwen3.8 27B on Hugging Face Inference Endpoints (Melvin Vivas on X · notes), Using Hugging Face credits: Jobs, Inference Endpoints and Open Models (Melvin Vivas on X · notes) and 38 more

Try this

• Open the Cohere Transcribe WebGPU Space and transcribe some audio locally in your browser.
• Build a private, in-browser transcription app on WebGPU using Cohere Transcribe.