LOADING DATASHEET
LOADING DATASHEET
by Hexgrad
RANKED #18 OF 31 AUDIO · OVERALL #357 OF 7,034 · VIBE SCORE 5.5 · 30 VOICES
Kokoro-82M-ONNX on Hugging Face (text to speech). 55,766 downloads. Open weights for local or hosted use.
8 mentions
2 mentions
26 weeks · 29 voices
AGING WELL
The crowd is warmer now than it was at the start.
Open weights. You can use a hosted service, or download it and run it yourself, free.
Open weights: download Kokoro 82M once and it's yours. No subscription, no rate limits, works offline.
82M parameters, FP32 file is about 330 MB and INT8/quantized ONNX is about 86 to 170 MB
Raspberry Pi 5
4 GB RAM
~$60 NEWONNX quantized INT8, realtime streaming on CPU, ~2x realtime speed
Mac mini M4
16 GB UNIFIED RAM
~$599 NEWFull FP32/FP16 precision, instant offline TTS pipeline, ~25x realtime speed
NVIDIA GeForce RTX 4070 SUPER
12 GB VRAM
~$599 NEWFull FP16 with CUDA / TensorRT acceleration, high-concurrency batch generation, ~120x realtime speed
The three cards above are researched picks. Search the machine you actually have — we only say yes if it should reply at a usable speed, not just load the weights and crawl.
HARDWARE PICKS & PRICES ARE RESEARCHED FROM THE LIVE WEB AND REFRESHED AUTOMATICALLY. TREAT THEM AS BALLPARK, NOT GOSPEL.
Everyday jobs on each of those rigs: how long Kokoro 82M takes, and how good the crowd says it is at that kind of work.
| THE JOB | BARE MINIMUMRaspberry Pi 5 | SWEET SPOTMac mini M4 | FULL POWERNVIDIA GeForce RTX 4070 SUPER |
|---|---|---|---|
| Narrate a scripta minute of spoken audio | ~30 S★★★★★ | ~2 S★★★★★ | UNDER A SECOND★★★★★ |
| Transcribe a recordinga ten minute recording turned into text | ~5 MIN★★★★★ | ~24 S★★★★★ | ~5 S★★★★★ |
STARS ARE THE COMMUNITY SCORE FOR THAT KIND OF WORK, DOCKED FOR HOW SQUASHED THE WEIGHTS GET AT EACH TIER. TIMES ARE COMPUTED FROM RESEARCHED SPEEDS. BALLPARK, NOT GOSPEL.
Not enough discussion about Kokoro 82M yet to call it. We found 30 posts but people did not say much either way.
28 POSTS · 2 COMMENTS · REDDIT · DEV.TO · BLUESKY · GITHUB
Other models the crowd has fully reviewed, starting with audio models like this one.
Kyutai's Pocket TTS clones a voice from 5 seconds of audio, on CPU, under MIT. Benchmarked against Kokoro, Supertonic, and Inflect-Nano for Eng. TTS
how to use kokoro with silly tavern in ubuntu
guide for kokoro v1.0 , now supports 8 languages, best TTS for low resources system(CPU and GPU)
Kokoro-82M on Ollama?
MimikaStudio - Voice Cloning, TTS & Audiobook Creator (macOS + Web): the most comprehensive open source app for voice cloning and TTS.
Working on a local TTS compatible PDF Reader (to use with Kokoro-82M)
Can anyone recommend a local open source TTS that has streaming and actual support for the GPU From a github project?
CPU TTS benchmark with UTMOS MOS scoring: Kokoro, Supertonic, Inflect-Nano, and Kyutai's new Pocket TTS [P]
Fool proof way to get Kokoro TTS (with streaming) working with SillyTavern
I made a DIY Claude Code voice assistant
Install Kokoro on PC
feat: add logus2k/tts_eu_pt European Portuguese Kokoro voice