Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#1

Post by xenova » Fri, Feb 07, 2025, 3:30 PM UTC

Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser
huggingface.co

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#2

Post by xenova » Fri, Feb 07, 2025, 3:30 PM UTC

It took some time, but we finally got Kokoro TTS (v1.0) running in-browser w/ WebGPU acceleration! This enables real-time text-to-speech without the need for a server. Looking forward to your feedback!

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#3

Post by waynenilsen » Fri, Feb 07, 2025, 4:00 PM UTC

Incredible work! I have listened to several tts and to have this be free and in complete control of the customer is absolutely incredible. This will unlock new use cases

I made https://app.readaloudto.me/ as a hobby thing and now it could be enhanced with a local tts option!

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#4

Post by rado » Fri, Feb 07, 2025, 4:09 PM UTC

Crashes the iPad Safari tab

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#5

Post by reach-vb » Fri, Feb 07, 2025, 4:41 PM UTC

Brilliant job! Love how fast it is, I'm sure if the rapid pace of speech ML continues we'll have Speech to Speech models directly running in our browser!

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#6

Post by zamadatix » Fri, Feb 07, 2025, 4:46 PM UTC

rado wrote: post #4

Crashes the iPad Safari tab

Mobile Safari (includes iPad) does not like to dish out large amounts of memory.

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#7

Post by dindresto » Fri, Feb 07, 2025, 4:51 PM UTC

zamadatix wrote: post #6

rado wrote: post #4

Crashes the iPad Safari tab

Mobile Safari (includes iPad) does not like to dish out large amounts of memory.

Same on macOS Safari (Sequoia, Safari 18.3, M3 Pro, 18gb RAM)

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#8

Post by fallinditch » Fri, Feb 07, 2025, 4:58 PM UTC

Brave browser and Samsung Galaxy S22 ultra - gives horrible screeching noises

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#9

Post by oliwary » Fri, Feb 07, 2025, 4:59 PM UTC

rado wrote: post #4

Crashes the iPad Safari tab

Worked on my Pixel 6a, albeit quite slowly (~30s for 4s audio). Still really impressed.

Re: Kokoro WebGPU: Real-time text-to-speech 100% locally in the browser

#10

Post by deivid » Fri, Feb 07, 2025, 5:00 PM UTC

xenova wrote: post #2

It took some time, but we finally got Kokoro TTS (v1.0) running in-browser w/ WebGPU acceleration! This enables real-time text-to-speech without the need for a server. Looking forward to your feedback!

Amazing! I'm interested in models running locally and Kokoro seems amazing. Are you aware of similar models but for Speech to text?