audiospeech
Kokoro
82M-parameter Kokoro TTS with 19 voices across 9 language endpoints.
Studio is in early access; availability resolves per project after sign-in.
- Endpoints
- 9
- Eligible
- 9
- Input schemas imported
- 9 / 9
- Delivery
- fal.ai
- Weights
- Not stated
- Developer
- Not stated in source
Overview
Kokoro TTS delivers natural speech synthesis on 82 million parameters, matching models 10-50x larger while running faster. Nineteen voices span 10 female and 9 male variants. Nine language endpoints (American and British English, Portuguese, French, Hindi, Italian, Japanese, Mandarin, Spanish) localize output.
Capabilities
- 82M efficient core
- 19 voice options
- 9 language endpoints
- 10-50x size advantage
Best for
- Voice agents
- Audiobooks
- Product localization
Use cases
- Content narration
- Lean deployment
- Multilingual synthesis
Endpoints
9 eligible. Cataloged is not the same as executable: Studio resolves which endpoints a project can run.
| Endpoint | Task | Catalog status | Inputs |
|---|---|---|---|
fal-ai/kokoro/american-english | text to audio | Eligible | 3 fields |
fal-ai/kokoro/brazilian-portuguese | text to audio | Eligible | 3 fields · requires prompt, voice |
fal-ai/kokoro/british-english | text to audio | Eligible | 3 fields · requires prompt, voice |
fal-ai/kokoro/french | text to audio | Eligible | 3 fields · requires prompt, voice |
fal-ai/kokoro/hindi | text to audio | Eligible | 3 fields · requires prompt, voice |
fal-ai/kokoro/italian | text to audio | Eligible | 3 fields · requires prompt, voice |
fal-ai/kokoro/japanese | text to audio | Eligible | 3 fields · requires prompt, voice |
fal-ai/kokoro/mandarin-chinese | text to audio | Eligible | 3 fields · requires prompt, voice |
fal-ai/kokoro/spanish | text to audio | Eligible | 3 fields · requires prompt, voice |
Questions
Model size?
82M parameters.
Voices?
19: 10 female, 9 male.