Skip to content
Agents tracked: 492 Releases (7d): 545 Agent status: 1 with issuesSince last limit reset: Claude 9d · Codex 7d For AI agents: Markdown Updated Nov 20, 2026

Kokoro-82M vs Chatterbox

Both are open-weight text-to-speech models you run yourself under permissive licenses: Apache 2.0 for Kokoro, MIT for Chatterbox. Kokoro is a single 82M-parameter model with a fixed set of preset voices, while Chatterbox clones a voice from a reference clip, covers 23 languages in its multilingual model and watermarks every file it generates.

Content asset vs content asset · checked against the official pages on Oct 25, 2026

Side by side

Kokoro-82MChatterbox
What it is82M-parameter open-weight TTS model under Apache-2.0 with fixed voice packs.MIT-licensed open TTS with zero-shot cloning in 23 languages and a built-in Perth watermark.
PriceFreeFree
Free tierSelf-hosted; compute only.Self-hosted; compute only (0.5B params).
LicenseApache-2.0MIT
Commercial useAllowedAllowed
How an agent connectsPackage (npm/PyPI)Package (npm/PyPI)
GitHub stars10.2K27.9K
Downloads, last 7 days258K91.9K
Gid Score3453

From AgentGid's catalogs: prices and licenses as read on the official pages, downloads and stars from npm, PyPI and GitHub, "used in agent projects" from our scan of public repositories set up for AI coding agents. Each item's page has its sources and dates.

Key differences

  • Kokoro-82MKokoro is one small open-weight model with 82 million parameters, run through the kokoro Python package. Source
  • Kokoro-82MKokoro's weights are Apache-licensed, and the card states it is already used in commercial APIs. Source
  • Kokoro-82MKokoro comes with a fixed list of preset voices, and some of its languages have only one or a few of them. Source
  • ChatterboxChatterbox Multilingual V3 is a 0.5B model that covers 23 languages. Source
  • ChatterboxChatterbox clones voices: its Turbo and Nano models generate speech from a reference recording of the target voice. Source
  • ChatterboxThe Chatterbox code is released under the MIT license. Source
  • ChatterboxEvery clip Chatterbox produces carries Resemble AI's imperceptible Perth watermark, which a provided script can detect. Source

Which one to pick

Pick Kokoro-82M if…Choose Kokoro when you want a very light speech model with ready-made voices that you can ship commercially under Apache 2.0. It suits English-first narration, apps and modest hardware where cloning a specific voice is not needed.
Pick Chatterbox if…Choose Chatterbox when you need to reproduce a particular voice or speak many languages with one MIT-licensed model. Expect a watermark in every output, and look at its smaller Turbo and Nano variants for English voice agents or CPU-only machines.

Sources

Vendor figures are as reported on their own pages; AgentGid did not test these products. Not legal advice: read the license before you ship.

Compare with another one → · All comparisons of libraries, APIs and assets

More comparisons

Apify vs Firecrawl · Browserbase vs Hyperbrowser · Browserbase vs Steel · Cartesia vs ElevenLabs · Clerk vs Auth0 · Deepgram vs AssemblyAI · E2B vs Daytona · ElevenLabs vs Kokoro-82M