Kokoro TTS
Hacker ApprovedTied for most recommended of 18 text-to-speech models
People recommend Kokoro for natural-sounding speech that runs locally on a CPU, is small, and is free and open-source. Common caveat: mistakes on some common phrases and numbers.
People recommend Kokoro for natural-sounding speech that runs locally on a CPU, is small, and is free and open-source. Common caveat: mistakes on some common phrases and numbers.
68 people gave an opinion
- recommend it
- 59
- criticize it
- 9
- approval share
- 87%
Recommended in 11 of 13 months
- November 2025: not recommended
- December 2025: recommended
- January 2026: recommended
- February 2026: recommended
- March 2026: recommended
- April 2026: recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Owned for 1 month
Median of the 3 owners who said how long they have had it.
Each square is one year.
Why it carries the badge
- 59 people recommend itNeeds 10 or more, met.
- 6.6 recommenders for every criticNeeds 2 to 1 or better, met.
- Recommended in 11 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Natural-sounding speech
“pretty good quality / natural sounding speech in mixed English / Japanese while running fast on CPU”
Comment from 7 Jul 2026 in “Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro” “kokoro which has the most natural for small size that I've seen”
Comment from 30 Aug 2026 in “Automating Immersive Reading” “The Kokoro voice is impressive for something generated locally.”
Comment from 30 Sep 2026 in “Show HN: Explainroo, Open Source Framework to Make Explainer Videos with Claude” - Comment from 20 Oct 2026 in “Voxtral 3”
- Comment from 7 Jul 2026 in “Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro”
- Comment from 19 Mar 2026 in “Show HN: Three new Kitten TTS models – smallest less than 25MB”
Runs fast on local CPU
“pretty good quality / natural sounding speech in mixed English / Japanese while running fast on CPU”
Comment from 7 Jul 2026 in “Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro” “Kokoro is a pretty popular tool- for good reason. Can run on CPUs on desktops and phone.”
Comment from 5 Apr 2026 in “We made significant improvements to the Kokoro TTS trainer” “My phone went from slightly slower than realtime to very comfortable”
Comment from 31 Aug 2026 in “Automating Immersive Reading” - Comment from 13 Oct 2026 in “Show HN: Use your OpenAI subscription to create audiobooks”
Small model size
“It's legitimately super good for being only 85mb (for the wasm version) or 300mb (for the webgpu version).”
Comment from 7 Jul 2026 in “Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro” “kokoro which has the most natural for small size that I've seen”
Comment from 30 Aug 2026 in “Automating Immersive Reading” “I find its the most natural sounding and small too.”
Comment from 7 Jul 2026 in “Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro” - Comment from 19 Mar 2026 in “Show HN: Three new Kitten TTS models – smallest less than 25MB”
Free and open-source
“Local, private, free to run, and enough voices that I can actually use a different one for each agent.”
Comment from 8 Jul 2026 in “Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro” “The voice layer (Parakeet + Kokoro) runs locally on the M4 so that's zero cost.”
Comment from 11 Feb 2026 in “Show HN: Mimora, a 3D avatar for OpenClaw AI agents with voice and expressions” “There are excellent free open-source TTS and STT models like KokoroTTS”
Comment from 27 Dec 2025 in “Free TTS and STT that runs 100% locally via WebGPU” - Comment from 8 Jul 2026 in “Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro”
What people warn about
Each mark is one comment
Mistakes on common phrases and numbers
“has a very good Norwegian voice but the rhythm and emphasizing is often so out of whack the generated speech is almost incomprehensible”
Comment from 19 Mar 2026 in “Show HN: Three new Kitten TTS models – smallest less than 25MB” “it's good overall, but it does make mistakes on some common phrases”
Comment from 25 Oct 2026 in “Dessa: Conversation Model trained only on speech” “that's why numbers, addresses, and hedged phrases sound off even when everything else works”
Comment from 16 Mar 2026 in “My Journey to a reliable and enjoyable locally hosted voice assistant” Slow or memory-heavy on some hardware
“it was too way too slow with latency also not supporting language I wanted”
Comment from 9 Jan 2026 in “Sopro TTS: A 169M model with zero-shot voice cloning that runs on the CPU” “It crashes as soon as you put a little paragraph of text on Apple M2 Pro.”
Comment from 8 Jul 2026 in “Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro” “Kokoro was great but memory was too high”
Comment from 22 Apr 2026 in “Show HN: Notation – an iOS chess coach (Stockfish and optional BYOK Claude)” - Comment from 30 Aug 2026 in “Automating Immersive Reading”
No streaming, lag between sentences
“Kokoro doesn't expose a streaming interface as far as I could find”
Comment from 15 Apr 2026 in “Show HN: I built on-device TTS app because I run out of audiobooks on a flight” “the issue was too much lag between sentences”
Comment from 19 Mar 2026 in “Show HN: Three new Kitten TTS models – smallest less than 25MB”
Other text-to-speech models people recommend
The four most recommended other text-to-speech models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Kokoro TTSThis page | 59 | 9 | 87% | 11 | 1 mo3 owners | Approved |
| ElevenLabs | 59 | 20 | 75% | 13 | — | Approved |
| Pocket TTS | 41 | 16 | 72% | 8 | — | Approved |
| Qwen3-TTS | 41 | 13 | 76% | 7 | — | Approved |
| Kitten TTS | 33 | 7 | 83% | 4 | — | Approved |
What people said
40 of the 83 comments that mention it, newest first. Each links to the original comment.
“it's good overall, but it does make mistakes on some common phrases”
“Best sounding local TTS engine I've heard by far is Kokoro”
“Kokoro is fast and I can run it on my own server”
“The Kokoro voice is impressive for something generated locally.”
“it outputs MP3, so I just copy those to my phone and listen as audiobooks”
“I've been making audiobooks out of text files that I have lying around, and kokoro TTS has been phenomenal”
“It helped me a lot to detect dialogues that don't flow naturally, or sentences that sounded awkward.”
“folks on the Latin language Discord said they got pretty good results using Kokoro TTS”
“My phone went from slightly slower than realtime to very comfortable”
“it was slower than real-time”
“kokoro which has the most natural for small size that I've seen”
“Works surprisingly well, as long as you are fine with consuming it”
“I built this over the weekend with Fable for a local voice chat running 100% on local LLMs, Parakeet and Kokoro”
“Kokoro has the added benefit of being open-source.”
“Local, private, free to run, and enough voices that I can actually use a different one for each agent.”
“I love Kokoro. I use it to read ebooks that don't have audiobooks and it works pretty good for that.”
“It crashes as soon as you put a little paragraph of text on Apple M2 Pro.”
“I replaced the TTS model with Kokoro after seeing this post, thank you.”
“pretty good quality / natural sounding speech in mixed English / Japanese while running fast on CPU”
“the male voices are all so much worse than the female voices (several are quite good)”
“I find its the most natural sounding and small too.”
“Both Text-to-Speech and Speech-to-Text now have local models that are good enough to get the job done.”
“we can get both rt natural local TTS and 4x less battery drainage and thermals”
“It's legitimately super good for being only 85mb (for the wasm version) or 300mb (for the webgpu version).”
“you could probably use Kokoro or some other TTS with a speed parameter”
“Kokoro was great but memory was too high”
“Kokoro doesn't expose a streaming interface as far as I could find”
“Kokoro is a pretty popular tool- for good reason. Can run on CPUs on desktops and phone.”
“the issue was too much lag between sentences”
“Kokoro is a really good model for its size”
“has a very good Norwegian voice but the rhythm and emphasizing is often so out of whack the generated speech is almost incomprehensible”
“that's why numbers, addresses, and hedged phrases sound off even when everything else works”
“I have a similar setup - I let it spin in the background on some various RSS news feeds that I can listen to while I go walking.”
“The voice layer (Parakeet + Kokoro) runs locally on the M4 so that's zero cost.”
“I've been a huge fan of Kokoro and event setup my own fully-local voice assistant”
“That's the stack, and not slow at all, but it depends on your HW.”
“it was too way too slow with latency also not supporting language I wanted”
“There are excellent free open-source TTS and STT models like KokoroTTS”
“self-hosted instance of Kokoro TTS (running on a Reserved VM) for high-quality, low-latency audio”
“even Kokoro-82M is faster and better than what we had before”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.