Loading video...

Video Failed to Load

Go Home

Introducing Voxtral WebGPU: Real-time speech transcription entirely in your browser. This demo runs Voxtral-Mini-4B, a powerful streaming ASR model from Mistral AI, locally on WebGPU. The model supports 13 languages and is capable of <500 ms latency. Fully private. Zero cost.

94,711 views • 8 months ago •via X (Twitter)

28 Comments

Xenova's profile picture
Xenova8 months ago

Powered by 🤗 Transformers.js v4 (preview). I'm excited to see what the community builds with it! Try it out yourself! Demo + source code 👇

Vlad's profile picture
Vlad8 months ago

@MistralAI this is fire, it switches languages just like that. Awesome!

Xenova's profile picture
Xenova8 months ago

@MistralAI Thanks! 🙏 I know right, the model is super impressive!

Vlad's profile picture
Vlad8 months ago

@MistralAI You have no idea how pump I am 😂 I feel like i have to add this somewhere to any of my free apps. 😂 Man the change from spanish to english and english to spanish right on the spot it's really insane.

DrFlask's profile picture
DrFlask8 months ago

@MistralAI

Xenova's profile picture
Xenova8 months ago

@MistralAI Which OS/browser are you running? We recommend Chrome (or a Chromium-based browser), and most of our tests are on Mac/Windows.

DrFlask's profile picture
DrFlask8 months ago

@MistralAI it could be that the lack of a standard graphics architecture on linux makes webgpu harder.

matt palmer's profile picture
matt palmer8 months ago

@MistralAI lets goooo!

Xenova's profile picture
Xenova8 months ago

@MistralAI You called it 😂

matt palmer's profile picture
matt palmer8 months ago

@MistralAI been following you long enough to know excited to build with this!

Haydon Ryan 🇦🇺🇺🇸's profile picture
Haydon Ryan 🇦🇺🇺🇸8 months ago

@MistralAI It’s good. Been running voxtral mini for a while with

Cesar Bonadio's profile picture
Cesar Bonadio8 months ago

@MistralAI Wohh, this is so good. It was able to correctly transcribe the name of my street, which is an Indigenous word in Portuguese. Congratulation!

岡田泰彦|光邦|AI×印刷エヴァンジェリスト's profile picture
岡田泰彦|光邦|AI×印刷エヴァンジェリスト8 months ago

@MistralAI I feel we're getting closer and closer to a world where advanced voice AI is readily available to everyone. 高度な音声AIが誰にでもすぐ使える世界に、どんどん近づいていると感じます。

GPUBenchTest.com's profile picture
GPUBenchTest.com8 months ago

@MistralAI Cannot wait till we drop our AI Performance results in our benchmark test 🥰 i can see the vision on this and will be curious to how Voxtral will compare to other loads Also can't wait to try them out

James's profile picture
James8 months ago

@MistralAI is this multithreaded? what kind of cpu/gpu usage does it have?

Alexandre Mendes's profile picture
Alexandre Mendes8 months ago

@MistralAI can i use it on a second window in youtube?

Jaga behera - NISM Certified's profile picture
Jaga behera - NISM Certified8 months ago

🚨 AI Tech Update Real-time speech transcription is now possible directly in your browser with Voxtral WebGPU. • Runs Voxtral‑Mini‑4B locally using WebGPU • Developed by Mistral AI • Supports 13 languages • <500 ms latency for near real-time transcription • Fully private processing happens on your device • Zero cost, no cloud required A glimpse into the future of on-device AI and private speech recognition. #AI #SpeechRecognition #WebGPU #MistralAI

LightShift Studio's profile picture
LightShift Studio8 months ago

@MistralAI 🎙️ Hey Insta fam!

richardanaya2_2048b.Q6_K.gguf 🇺🇸🤖's profile picture
richardanaya2_2048b.Q6_K.gguf 🇺🇸🤖8 months ago

@MistralAI Does this run on Oculus Quest chrome browser?

0xShugo's profile picture
0xShugo4 months ago

@MistralAI C++/GGML based Voxtral Mini 4B Realtime released! 16.39x (0.0610 RTF) realtime offline, 170ms TTFT streaming on RTX 5090.

Ariel Jankelowitz's profile picture
Ariel Jankelowitz8 months ago

@MistralAI Amazing, can’t wait to integrate this.

Melvin Vivas's profile picture
Melvin Vivas8 months ago

@MistralAI wow, must try this

Koshur AI's profile picture
Koshur AI8 months ago

@MistralAI

kaiyes's profile picture
kaiyes8 months ago

@MistralAI this should be able to transcribe youtube videos right ?

Kim Hjortholm's profile picture
Kim Hjortholm8 months ago

@MistralAI Argh, no support for Danish - my test got transmogrified to Dutch 😂 Which actually makes sense as the best guess in the lack of Danish support

kfant's profile picture
kfant8 months ago

@MistralAI yea but new granite by ibm ...

Faust's profile picture
Faust8 months ago

@MistralAI Been waiting for something like this. Running a 4B model locally with sub-500ms latency is wild—WebGPU keeps unlocking new possibilities.

Victor's profile picture
Victor8 months ago

@Scobleizer @MistralAI lol the fact that this runs entirely in browser with zero backend is kind of insane, been waiting for local speech-to-text that actually works 😭

Related Videos