Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro

ariya.io

Post reply on HN
Print view View original

101–107 of 107 posts

Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro

#101

Post by daqhris »

Kokoro voices were the non-human actors directed by Opus, featured in the sound performance "God Bless the United States of Aliens". A theatrical adaptation of a liturgical texts, a surveilled mass ceremony at a protesting Christian church, being exhibited at the 61st Venice Biennale until November 2026 (Solidarity Burundi Pavilion, inside of Bosnia-Herzegovina Pavilion).

The various models of Kokoro TTS were theatrically conducted by Claude during an Anthropic hackathon. AI Voices were generated locally on a low-spec laptop to save up on API tokens and stretch the hackathon budget. Kokoro has the added benefit of being open-source.

https://daqhris.com/venice-biennale

https://daqhris.com/god-bless-usa/submission.html

Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro

#102

Post by fivestones »

mackross wrote: post #90

Supertonic 3 is pretty amazing too. I’ve got time to first sound down to well under a second by streaming chunks and stitching together from CPU with goonx (non c-go onnx) - happy to upload model if any go devs interested.

Interested

Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro

#104

Post by jasonm23 »

dmayle wrote: post #15
dghlsakjg wrote: post #14

Earlier quoted context omitted.

This is TTS. Not STT. For what you are doing, Senko works really well for diarization along with parakeet. Faster and more accurate than Pyannote and whisper on my MacBook anyway.

You're right... I read the title too quickly... I'll have to look at Senko vs Softformer later...

Still I for one am grateful for any STT transcription models.

Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro

#105

Post by mowmiatlas »

tigerquoll wrote: post #36
mowmiatlas wrote: post #16

Cool I actually got it ported to iPhone’s ANE finally yesterday! So we can get both rt natural local TTS and 4x less battery drainage and thermals

lol, I just finished a port to the iPhone ANE as well. I would love to compare notes

The biggest part was dynamic shapes that weren’t allowed for newer iOS. I’d do it with masking similar to other transformers being ported to ANE. Another thing was people would cut the graph cause some operations are not allowable on this chip - tho if you look at the op that is doing segfault, i.e. tile - you can rewrite the same functionality with other supported ops and port the weights to the new graph. Works the same way compared to PyTorch weights comparing spectrograms, around 10x realtime on 14pro :)

Anything you found interesting there? Every proposal on GitHub I found was kinda lucking in both documentation and completeness

Post reply on HN