Kitten TTS

New: Free Kitten TTS API available at https://platform.kittenml.com

Kitten TTS is an open-source text-to-speech library. Its flagship model, KittenTTS 2, is a 1.7B-parameter speech language model with in-context voice cloning and expression control: give it five seconds of anyone's voice and it speaks your text in that voice.

The library also ships the original lightweight legacy models, 15M-80M parameters, which run on CPU without a GPU. Both families load through the same KittenTTS(...) constructor.

Commercial support is available. For integration assistance, custom voices, or enterprise licensing, contact us.

Table of Contents

• Features
• Available Models
• Demo
• Quick Start
• Voice cloning
• Expression controls
• Long text and streaming
• Documentation
• System Requirements
• Roadmap
• Commercial Support
• Community and Support
• License

Features

• Voice cloning -- Clone any speaker from 5-30 seconds of audio, no fine-tuning
• 47 built-in voices -- Including the eight from KittenTTS 0.8 and nine non-English
• Multilingual -- Ten languages: English, Arabic, Chinese, French, German, Hindi, Italian, Portuguese, Russian, Spanish
• Expression control -- [emotion] tags, inline <event> tags, and (((emphasis))) spans
• Decoding presets -- Trade stability against expressiveness per request
• Long-form text -- Sentence-aware chunking with seamless joins
• Text preprocessing -- Numbers, currencies, dates, units and abbreviations expanded automatically
• 24 kHz output -- High-quality audio at a standard sample rate
• Runs without a GPU

Available Models

KittenTTS 2 -- speech language model, GPU recommended:

Model	Parameters	Size	Voices	Download
kitten-tts-2	1.7B	947 MiB	47 + cloning	KittenML/kitten-tts-2

Lightweight ONNX -- runs on CPU, no GPU required:

Model	Parameters	Size	Download
kitten-tts-mini	80M	80 MB	KittenML/kitten-tts-mini-0.8
kitten-tts-micro	40M	41 MB	KittenML/kitten-tts-micro-0.8
kitten-tts-nano	15M	56 MB	KittenML/kitten-tts-nano-0.8
kitten-tts-nano (int8)	15M	25 MB	KittenML/kitten-tts-nano-0.8-int8

Note: Some users have reported issues with the kitten-tts-nano-0.8-int8 model. If you encounter problems, please open an issue.

Demo

https://github.com/user-attachments/assets/d80120f2-c751-407e-a166-068dd1dd9e8d

Try it online

Try Kitten TTS directly in your browser on Hugging Face Spaces.

Quick Start

Prerequisites

• Python 3.9 or later
• A CUDA GPU with roughly 8 GB free, or a CPU with about 6 GB of RAM
• About 1 GB of disk space for the model

Installation

pip install kittenml

That is the whole install. KittenTTS 2 is the default model, voice cloning is included, and no Hugging Face login is needed -- every weight the model uses ships in its own repository. It also pulls the small ONNX runtime, so the lightweight models work from the same install.

Basic usage

from kittenml import KittenTTS
import soundfile as sf

m = KittenTTS("KittenML/kitten-tts-2")

audio = m.generate("One day, a little girl named Lily found a needle in her room.",
                   voice="Bruno")
sf.write("output.wav", audio, m.sample_rate)

m.available_voices lists all 47 built-in voices, described in voices and expression. Bella, Jasper, Luna, Bruno, Rosie, Hugo, Kiki and Leo are the same speakers as in KittenTTS 0.8, so code written against the ONNX models keeps working.

# Trade stability against expressiveness
audio = m.generate("Hello, world.", voice="Luna", preset="expressive")

# Save directly to a file
m.generate_to_file("Hello, world.", "output.wav", voice="Bruno")

Voice cloning

Pass reference= instead of voice= — same method, the recording just replaces the built-in speaker. Give it 5-30 seconds of a single speaker. The transcript is part of the prompt, but you do not have to type it; Whisper fills it in when omitted.

audio = m.generate("This is my own voice, cloned.", reference="my_voice.wav")

# Supplying the transcript skips the Whisper pass
audio = m.generate("This is my own voice.", reference="my_voice.wav",
                   reference_text="what is actually said in the clip")

The reference feeds the model by two independent routes -- a speaker embedding through the model's projection head, and the clip itself as codec tokens in the prompt -- so identity survives even when one route is weak.

Measured on the built-in voices: a generated clip scores 0.49-0.72 speaker similarity against its own reference and 0.01-0.18 against the other 37, and cloning an unseen recording scores 0.81 against that recording.

Expression controls

audio = m.generate(
    "[joyful] We actually won the grant <laugh> I can (((hardly))) believe it!",
    voice="Kiki",
    preset="expressive",
)

A leading [emotion] tag, inline <event> tags and (((emphasis))) spans reach the model as markup rather than being spoken, and automatically enable its expression conditioning. Ten emotions and ten vocal events are recognised -- see voices and expression for the full lists and what is not covered.

Long text and streaming

Long input is split on sentence boundaries and synthesized chunk by chunk, then joined with silence trimming and short edge fades so the seams are inaudible. This is automatic: the model is reliable on short inputs but truncates or drifts into repetition when asked for a whole script in one pass.

To start playing before the whole thing is ready, stream it:

for chunk in m.generate_stream(long_text, voice="Luna"):
    play(chunk)          # each chunk is a numpy array at m.sample_rate

It takes the same arguments as generate, so reference= streams a cloned voice too.

Streaming is chunk-level, not token-level: a chunk is generated and vocoded in full before it is yielded, so the first chunk still costs its own generation time. On an A100, a 936-character passage yielded its first 21 s of audio after 17 s and finished 54 s of audio in 42 s of wall clock -- so playback keeps ahead of generation, but there is a real initial delay.

Two consequences worth knowing:

• Short text does not stream. Input that fits in one chunk (under roughly 380 characters, and short trailing pieces get merged into their neighbour) yields exactly one chunk, so generate_stream behaves like generate.
• Chunks are yielded raw. generate post-processes the seams -- trimming each segment's edge silence, adding short fades and one consistent pause -- which a streaming caller cannot do without waiting for the next chunk. Concatenating streamed chunks directly gives slightly rougher joins than generate on the same text.

Documentation

API reference	Every argument to generate, streaming, and the advanced knobs
Voices and expression	The 47 voices, emotion and vocal-event tags, the ten languages
Decoders	How audio is decoded, and the smaller quantised decoders
Text normalization	How written text becomes spoken text
Architecture	What the model is, package layout, vendored components
Lightweight ONNX models	The CPU models, 15M-80M parameters, and their API

System Requirements

KittenTTS 2

• Operating system: Linux, Windows or Mac
• Python: 3.9 or later

Lightweight legacy models

• Operating system: Linux, macOS, or Windows
• Python: 3.8 or later
• Hardware: Runs on CPU; no GPU required

A virtual environment (conda, venv, or similar) is recommended to avoid dependency conflicts.

Commercial Support

We offer commercial support for teams integrating Kitten TTS into their products. This includes integration assistance, custom voice development, and enterprise licensing.

Contact us or email info@stellonlabs.com to discuss your requirements.

Community and Support

• Discord: Join the community
• Website: kittenml.com
• Custom support: Request form
• Email: info@stellonlabs.com
• Issues: GitHub Issues

License

This project is licensed under the Apache License 2.0.