# hexgrad/kokoro

https://hf.co/hexgrad/Kokoro-82M

Repository: https://github.com/hexgrad/kokoro
Canonical: https://ross.abutalabs.com/products/kokoro
Language: JavaScript
License: Apache-2.0
License Family: permissive
Last push: 2025-08-06T22:28:53+00:00

## Health v2 (maintenance only)
Score: 36/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 35, release rhythm 35, longevity 42
- inputs: {"age_days": 600, "days_push": 392, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 8570, forks 932 (observed 2026-08-28T04:10:23.589085+00:00)

## What it is
An inference library for Kokoro-82M, an open-weight text-to-speech model with 82 million parameters that delivers quality comparable to larger models while being faster and cheaper. It provides a Python pipeline (KPipeline) supporting multiple languages and voices, built on the misaki G2P library.

## Use cases
- convert text to natural-sounding speech in python
- generate voice audio for videos or podcasts
- run a lightweight tts model on cpu or gpu
- synthesize speech in multiple languages like english, japanese, or chinese
- add narration to an application with an apache-licensed tts model
- produce wav files from text offline

## When to choose
- you need fast, high-quality tts with a small model footprint
- you want permissively licensed (apache-2.0) voice synthesis for commercial use
- you need multilingual speech generation from a single pipeline

## When to avoid
- you need voice cloning or custom voice training
- you require streaming real-time tts with sub-100ms latency
- you need a non-python runtime without extra integration work

## Facets
- artifact type: library
- maturity: active
- function: tts, speech-recognition, machine-learning, llm-inference
- domain: speech-processing, artificial-intelligence
- platform: python, cross-platform
- tags: text-to-speech, voice-synthesis, lightweight-model, g2p, open-weights, audio, gpu

## Member repositories
- hexgrad/kokoro (main) score 36

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:23.589085+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:26:17.133374+00:00, confidence not recorded.
  - readme: https://github.com/hexgrad/kokoro (fetched 2026-08-28T04:10:23.589085+00:00, sha 01577210933e)
- Data as of 2026-08-30T08:39:29.467469+00:00.