Skip to page content
SAVRN
Search Plan a project

Open-weight model · Text to speech

pocket-tts-onnx

by Nitesh Sharma theoracleguy/pocket-tts-onnx

pocket-tts-onnx is an open-weight model for text to speech from Nitesh Sharma, released under Creative Commons Attribution 4.0. Its published files total 12.1 GB.

Optimized ONNX exports of Pocket TTS for high-speed, memory-efficient text-to-speech and zero-shot voice cloning on CPU, mobile, and edge runtimes. Mirrored and validated from KevinAHM/pocket-tts-onnx at commit 58a6d00cf13d239b6748cb0769f35c580a8f606c.

Parameters—
Context—
Weights12.1 GB
Licensecc-by-4.0
AccessOpen weights
Monthly Downloads—

Model Card

By Nitesh Sharma, published under cc-by-4.0, revision 9efb2a5306d9.

Optimized ONNX exports of Pocket TTS for high-speed, memory-efficient text-to-speech and zero-shot voice cloning on CPU, mobile, and edge runtimes. Mirrored and validated from KevinAHM/pocket-tts-onnx at commit 58a6d00cf13d239b6748cb0769f35c580a8f606c. This repository provides comprehensive support for both compact 6-layer models and high-capacity 24-layer models: - 6-Layer Compact Models: english2026-04, german, italian, portuguese, spanish - Ideal for lightweight CPU inference, mobile devices, and low-latency streaming. - ~302 MB FP32 / ~76 MB INT8 Flow LM. - 24-Layer Capacity Models: french24l, german24l, italian24l, portuguese24l, spanish24l - High-fidelity modeling for complex prosody…

Read Nitesh Sharma's full model card

Pocket TTS ONNX (Complete Multilingual & 24L Export)

Optimized ONNX exports of Pocket TTS for high-speed, memory-efficient text-to-speech and zero-shot voice cloning on CPU, mobile, and edge runtimes.

Mirrored and validated from KevinAHM/pocket-tts-onnx at commit 58a6d00cf13d239b6748cb0769f35c580a8f606c.

Available Language Bundles

This repository provides comprehensive support for both compact 6-layer models and high-capacity 24-layer models:

  • 6-Layer Compact Models: english_2026-04, german, italian, portuguese, spanish
  • Ideal for lightweight CPU inference, mobile devices, and low-latency streaming.
  • ~302 MB FP32 / ~76 MB INT8 Flow LM.
  • 24-Layer Capacity Models: french_24l, german_24l, italian_24l, portuguese_24l, spanish_24l
  • High-fidelity modeling for complex prosody, including full French language support.
  • ~1.21 GB FP32 / ~305 MB INT8 Flow LM.

Artifacts Per Language Bundle

Each bundle directory in onnx/<language>/ provides complete standalone inference:

  • flow_lm_main.onnx & flow_lm_main_int8.onnx: Dual-split conditioning transformer
  • flow_lm_flow.onnx & flow_lm_flow_int8.onnx: Flow matching ODE step model
  • mimi_decoder.onnx & mimi_decoder_int8.onnx: Neural audio codec decoder
  • mimi_encoder.onnx & mimi_encoder_int8.onnx: Voice prompt audio encoder
  • text_conditioner.onnx & text_conditioner_int8.onnx: Text embedding conditioner
  • bos_before_voice.npy: Beginning-of-sequence voice prompt vector (4,224 bytes)
  • tokenizer.model: Language-specific SentencePiece model
  • bundle.json: Bundle configuration and predefined voice registry
  • download_manifest.json: Per-bundle verification manifest with byte sizes and SHA-256 checksums

Root-level convenience models under onnx/ are retained for backward compatibility with flat loaders.

Quickstart

Installation

pip install -r requirements.txt

Python Inference

from pocket_tts_onnx import PocketTTSOnnx

# Load English 6L bundle (INT8 quantization by default)
tts_en = PocketTTSOnnx(models_dir="onnx", language="english_2026-04", precision="int8")
audio = tts_en.generate("Hello world, this is a test of Pocket TTS ONNX.", voice="alba")
tts_en.save_audio(audio, "output_en.wav")

# Load French 24L bundle
tts_fr = PocketTTSOnnx(models_dir="onnx", language="french_24l", precision="int8")
audio_fr = tts_fr.generate("Bonjour tout le monde, ceci est un test.", voice="alba")
tts_fr.save_audio(audio_fr, "output_fr.wav")

CLI Synthesis

python generate.py "Hello from Pocket TTS ONNX." alba output.wav --language english_2026-04
python generate.py "Bonjour le monde." alba output_fr.wav --language french_24l

Integrity & Verification

A root repository_manifest.json lists every published artifact with its precise byte size and SHA-256 checksum for automated verification.

License & Attribution

The Pocket TTS model weights and architecture are developed by Kyutai and licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). Upstream ONNX conversions by KevinAHM.

Identity and Version

Repository
theoracleguy/pocket-tts-onnx
Publisher
Nitesh Sharma
Task
Text to speech
Modality
Audio
Library
pocket-tts-onnx
Parameters
Not stated by the source
Languages
de, en, es, fr, it, pt
Revision
9efb2a5306d9f15056be42da3840081f94128f64
First published
2026-09-14
Last updated
2026-09-21

Files and Weights

157 files, 12.1 GB in total. The weights are 108 files totalling 12.1 GB in onnx.

Weights108 files · 12.1 GB
Configuration23 files · 415.3 KB
Tokenizer11 files · 662.5 KB
Documentation2 files · 22.1 KB
Other12 files · 328.4 KB
Repository1 file · 1.6 KB
Every file
FileTypeSizeSHA-256
onnx/english_2026-04/flow_lm_flow.onnxWeights39.1 MB 085d239f6889
onnx/english_2026-04/flow_lm_flow_int8.onnxWeights10.0 MB 3dd781ee5abe
onnx/english_2026-04/flow_lm_main.onnxWeights302.7 MB 6d18315e2c33
onnx/english_2026-04/flow_lm_main_int8.onnxWeights76.3 MB f9bd8106b79a
onnx/english_2026-04/mimi_decoder.onnxWeights41.5 MB 86f038caa02a
onnx/english_2026-04/mimi_decoder_int8.onnxWeights22.7 MB 3630450a3297
onnx/english_2026-04/mimi_encoder.onnxWeights39.8 MB 853e2ca623b8
onnx/english_2026-04/mimi_encoder_int8.onnxWeights20.8 MB abb46f1ad0b8
onnx/english_2026-04/text_conditioner.onnxWeights16.4 MB 4ecee995fb69
onnx/english_2026-04/text_conditioner_int8.onnxWeights16.4 MB 24c0642bff11
onnx/flow_lm_flow.onnxWeights39.1 MB cf738ec75928
onnx/flow_lm_flow_int8.onnxWeights10.0 MB 8d627d235c44
onnx/flow_lm_main.onnxWeights302.7 MB 1c5f40935cd4
onnx/flow_lm_main_int8.onnxWeights76.3 MB fd5cdd7f7ab0
onnx/french_24l/flow_lm_flow.onnxWeights39.1 MB 0a098e70f484
onnx/french_24l/flow_lm_flow_int8.onnxWeights10.0 MB d340c549d5a1
onnx/french_24l/flow_lm_main.onnxWeights1.2 GB 4cf6b6571065
onnx/french_24l/flow_lm_main_int8.onnxWeights305.1 MB 6130a6b98fae
onnx/french_24l/mimi_decoder.onnxWeights41.5 MB 7c46912a3e2b
onnx/french_24l/mimi_decoder_int8.onnxWeights22.7 MB b329ff3de3aa
onnx/french_24l/mimi_encoder.onnxWeights39.8 MB 790a82fb8041
onnx/french_24l/mimi_encoder_int8.onnxWeights20.8 MB f661fcb2dbf3
onnx/french_24l/text_conditioner.onnxWeights16.4 MB 89d1b6ac55e6
onnx/french_24l/text_conditioner_int8.onnxWeights16.4 MB bc74572507e0
onnx/german/flow_lm_flow.onnxWeights39.1 MB 3b6dceddb6ea
onnx/german/flow_lm_flow_int8.onnxWeights10.0 MB 2b0eb8381c49
onnx/german/flow_lm_main.onnxWeights302.7 MB 368c05f4b275
onnx/german/flow_lm_main_int8.onnxWeights76.3 MB 3d5e9cd303ec
onnx/german/mimi_decoder.onnxWeights41.5 MB e068e4b2a30e
onnx/german/mimi_decoder_int8.onnxWeights22.7 MB 3fe868810d66
onnx/german/mimi_encoder.onnxWeights39.8 MB 71a00b2c6270
onnx/german/mimi_encoder_int8.onnxWeights20.8 MB d11af63a34c0
onnx/german/text_conditioner.onnxWeights16.4 MB a004c74047bc
onnx/german/text_conditioner_int8.onnxWeights16.4 MB de27b06fcd39
onnx/german_24l/flow_lm_flow.onnxWeights39.1 MB 3b6dceddb6ea
onnx/german_24l/flow_lm_flow_int8.onnxWeights10.0 MB 2b0eb8381c49
onnx/german_24l/flow_lm_main.onnxWeights1.2 GB 7ee0cf9bdae4
onnx/german_24l/flow_lm_main_int8.onnxWeights305.1 MB 3d9661a01a08
onnx/german_24l/mimi_decoder.onnxWeights41.5 MB e068e4b2a30e
onnx/german_24l/mimi_decoder_int8.onnxWeights22.7 MB 3fe868810d66
onnx/german_24l/mimi_encoder.onnxWeights39.8 MB 7367f8af75d7
onnx/german_24l/mimi_encoder_int8.onnxWeights20.8 MB 9c235c292b5e
onnx/german_24l/text_conditioner.onnxWeights16.4 MB e9e939c3da9b
onnx/german_24l/text_conditioner_int8.onnxWeights16.4 MB e1e93583d26d
onnx/italian/flow_lm_flow.onnxWeights39.1 MB 3646665d2b64
onnx/italian/flow_lm_flow_int8.onnxWeights10.0 MB 21b2bec2f9ae
onnx/italian/flow_lm_main.onnxWeights302.7 MB 43e83dd16d52
onnx/italian/flow_lm_main_int8.onnxWeights76.3 MB f43ce4d82347
onnx/italian/mimi_decoder.onnxWeights41.5 MB f474670d1494
onnx/italian/mimi_decoder_int8.onnxWeights22.7 MB f120bc5cddca
onnx/italian/mimi_encoder.onnxWeights39.8 MB 8936e1f95bae
onnx/italian/mimi_encoder_int8.onnxWeights20.8 MB 987796434789
onnx/italian/text_conditioner.onnxWeights16.4 MB 692369f5ac34
onnx/italian/text_conditioner_int8.onnxWeights16.4 MB 5cbccc5923b4
onnx/italian_24l/flow_lm_flow.onnxWeights39.1 MB 3646665d2b64
onnx/italian_24l/flow_lm_flow_int8.onnxWeights10.0 MB 21b2bec2f9ae
onnx/italian_24l/flow_lm_main.onnxWeights1.2 GB e213ce648733
onnx/italian_24l/flow_lm_main_int8.onnxWeights305.1 MB 8c7b0dc076e6
onnx/italian_24l/mimi_decoder.onnxWeights41.5 MB f474670d1494
onnx/italian_24l/mimi_decoder_int8.onnxWeights22.7 MB f120bc5cddca
onnx/italian_24l/mimi_encoder.onnxWeights39.8 MB 1f436e797ffd
onnx/italian_24l/mimi_encoder_int8.onnxWeights20.8 MB 9f45bdf936a6
onnx/italian_24l/text_conditioner.onnxWeights16.4 MB 00ff8327e832
onnx/italian_24l/text_conditioner_int8.onnxWeights16.4 MB e3113c92b9e8
onnx/mimi_decoder.onnxWeights41.5 MB f58ecc92a349
onnx/mimi_decoder_int8.onnxWeights22.7 MB 501e16f51cf3
onnx/mimi_encoder.onnxWeights73.2 MB 360f050cd0b1
onnx/portuguese/flow_lm_flow.onnxWeights39.1 MB 41134866b54a
onnx/portuguese/flow_lm_flow_int8.onnxWeights10.0 MB 53c50fc75f89
onnx/portuguese/flow_lm_main.onnxWeights302.7 MB 56c3d0c36583
onnx/portuguese/flow_lm_main_int8.onnxWeights76.3 MB 119e5ce926ca
onnx/portuguese/mimi_decoder.onnxWeights41.5 MB e396510a578a
onnx/portuguese/mimi_decoder_int8.onnxWeights22.7 MB 691f89f07458
onnx/portuguese/mimi_encoder.onnxWeights39.8 MB 189d737f150a
onnx/portuguese/mimi_encoder_int8.onnxWeights20.8 MB 37029c2653a2
onnx/portuguese/text_conditioner.onnxWeights16.4 MB 693523a9b8e8
onnx/portuguese/text_conditioner_int8.onnxWeights16.4 MB 0e046165bd63
onnx/portuguese_24l/flow_lm_flow.onnxWeights39.1 MB 41134866b54a
onnx/portuguese_24l/flow_lm_flow_int8.onnxWeights10.0 MB 53c50fc75f89
onnx/portuguese_24l/flow_lm_main.onnxWeights1.2 GB 831fb01efdcd
onnx/portuguese_24l/flow_lm_main_int8.onnxWeights305.1 MB 6706505e9a57
onnx/portuguese_24l/mimi_decoder.onnxWeights41.5 MB e396510a578a
onnx/portuguese_24l/mimi_decoder_int8.onnxWeights22.7 MB 691f89f07458
onnx/portuguese_24l/mimi_encoder.onnxWeights39.8 MB 55621c0227fe
onnx/portuguese_24l/mimi_encoder_int8.onnxWeights20.8 MB 8a63c79d5233
onnx/portuguese_24l/text_conditioner.onnxWeights16.4 MB e6aa04d1905b
onnx/portuguese_24l/text_conditioner_int8.onnxWeights16.4 MB 54b8eae9916d
onnx/spanish/flow_lm_flow.onnxWeights39.1 MB 2f6300c6a07e
onnx/spanish/flow_lm_flow_int8.onnxWeights10.0 MB 4cc630a8fc1d
onnx/spanish/flow_lm_main.onnxWeights302.7 MB 45c1a49d6c8b
onnx/spanish/flow_lm_main_int8.onnxWeights76.3 MB 9b020318976e
onnx/spanish/mimi_decoder.onnxWeights41.5 MB ee9684cd12e6
onnx/spanish/mimi_decoder_int8.onnxWeights22.7 MB 309efd11b65c
onnx/spanish/mimi_encoder.onnxWeights39.8 MB 6702fb117c58
onnx/spanish/mimi_encoder_int8.onnxWeights20.8 MB c2d5b57c7acb
onnx/spanish/text_conditioner.onnxWeights16.4 MB 836ad5231fff
onnx/spanish/text_conditioner_int8.onnxWeights16.4 MB d6b0f7600ded
onnx/spanish_24l/flow_lm_flow.onnxWeights39.1 MB 2f6300c6a07e
onnx/spanish_24l/flow_lm_flow_int8.onnxWeights10.0 MB 4cc630a8fc1d
onnx/spanish_24l/flow_lm_main.onnxWeights1.2 GB 6392d64574aa
onnx/spanish_24l/flow_lm_main_int8.onnxWeights305.1 MB a37cc7c0ad34
onnx/spanish_24l/mimi_decoder.onnxWeights41.5 MB ee9684cd12e6
onnx/spanish_24l/mimi_decoder_int8.onnxWeights22.7 MB 309efd11b65c
onnx/spanish_24l/mimi_encoder.onnxWeights39.8 MB 7dc26cedb4d5
onnx/spanish_24l/mimi_encoder_int8.onnxWeights20.8 MB 146aaa643a59
onnx/spanish_24l/text_conditioner.onnxWeights16.4 MB 96a0e96e1f0a
onnx/spanish_24l/text_conditioner_int8.onnxWeights16.4 MB e1892e08e6cd
onnx/text_conditioner.onnxWeights16.4 MB 80ea69f46d81
generate.pyConfiguration2.1 KB —
onnx/english_2026-04/bundle.jsonConfiguration24.4 KB —
onnx/english_2026-04/download_manifest.jsonConfiguration2.4 KB —
onnx/french_24l/bundle.jsonConfiguration42.2 KB —
onnx/french_24l/download_manifest.jsonConfiguration2.3 KB —
onnx/german/bundle.jsonConfiguration24.4 KB —
onnx/german/download_manifest.jsonConfiguration2.2 KB —
onnx/german_24l/bundle.jsonConfiguration42.2 KB —
onnx/german_24l/download_manifest.jsonConfiguration2.3 KB —
onnx/italian/bundle.jsonConfiguration24.4 KB —
onnx/italian/download_manifest.jsonConfiguration2.3 KB —
onnx/italian_24l/bundle.jsonConfiguration42.2 KB —
onnx/italian_24l/download_manifest.jsonConfiguration2.3 KB —
onnx/portuguese/bundle.jsonConfiguration24.4 KB —
onnx/portuguese/download_manifest.jsonConfiguration2.3 KB —
onnx/portuguese_24l/bundle.jsonConfiguration42.2 KB —
onnx/portuguese_24l/download_manifest.jsonConfiguration2.4 KB —
onnx/spanish/bundle.jsonConfiguration24.4 KB —
onnx/spanish/download_manifest.jsonConfiguration2.3 KB —
onnx/spanish_24l/bundle.jsonConfiguration42.2 KB —
onnx/spanish_24l/download_manifest.jsonConfiguration2.3 KB —
pocket_tts_onnx.pyConfiguration28.8 KB —
repository_manifest.jsonConfiguration28.3 KB —
README.mdDocumentation3.5 KB —
onnx/LICENSEDocumentation18.7 KB —
onnx/english_2026-04/bos_before_voice.npyOther4.2 KB f46edf4f7007
onnx/french_24l/bos_before_voice.npyOther4.2 KB 5759269802d6
onnx/german/bos_before_voice.npyOther4.2 KB 8c866aec4bcd
onnx/german_24l/bos_before_voice.npyOther4.2 KB 51c947a558eb
onnx/italian/bos_before_voice.npyOther4.2 KB 212357ca66b4
onnx/italian_24l/bos_before_voice.npyOther4.2 KB deced1318865
onnx/portuguese/bos_before_voice.npyOther4.2 KB 9396378d581d
onnx/portuguese_24l/bos_before_voice.npyOther4.2 KB 449db3e61656
onnx/spanish/bos_before_voice.npyOther4.2 KB 2c77a716ac20
onnx/spanish_24l/bos_before_voice.npyOther4.2 KB 74e3cf820cb0
reference_sample.wavOther286.1 KB 88fbb0d31ec2
requirements.txtOther84 B —
.gitattributesRepository1.6 KB —
onnx/english_2026-04/tokenizer.modelTokenizer59.3 KB d461765ae179
onnx/french_24l/tokenizer.modelTokenizer60.2 KB 521c85bdb2da
onnx/german/tokenizer.modelTokenizer59.8 KB 389079b9c67c
onnx/german_24l/tokenizer.modelTokenizer59.8 KB 389079b9c67c
onnx/italian/tokenizer.modelTokenizer60.1 KB 6583b974a11b
onnx/italian_24l/tokenizer.modelTokenizer60.1 KB 6583b974a11b
onnx/portuguese/tokenizer.modelTokenizer61.0 KB 3aa51309c55f
onnx/portuguese_24l/tokenizer.modelTokenizer61.0 KB 3aa51309c55f
onnx/spanish/tokenizer.modelTokenizer60.9 KB aac2b96478e3
onnx/spanish_24l/tokenizer.modelTokenizer60.9 KB aac2b96478e3
tokenizer.modelTokenizer59.3 KB d461765ae179

License and Download

License
cc-by-4.0
Access
Open weights, no gate
Download size
12.1 GB
Download from Nitesh Sharma

Released by Nitesh Sharma through its official repository on Hugging Face. Read the license.

Built From

  • Derived from kyutai/pocket-tts
  • Quantized from kyutai/pocket-tts

Memory Requirements

PrecisionWeights in memory
As published12.1 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About pocket-tts-onnx

Can I use pocket-tts-onnx commercially?

Yes. pocket-tts-onnx is released under Creative Commons Attribution 4.0. CC BY 4.0 permits sharing and adapting the work, including commercially, provided the creator is credited and changes are indicated.

Similar Models

Model · Text to speech

Kokoro-82M

Hexgrad

Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers comparable quality to larger models while being significantly faster and more cost-efficient. With Apache-licensed weights, Kokoro can be deployed anywhere from production environments to personal projects. You can run this basic cell on Google Colab. Listen to samples. For more languages and details, see Advanced Usage. Under the hood, kokoro uses misaki, a G2P library at https://github.com/hexgrad/misaki Model SHA256 Hash: 496dba118d1a58f5f3db2efc88dbdc216e0483fc89fe6e47ee1f2c53f18ad1e4 Data: Kokoro was trained exclusively on permissive/non-copyrighted audio data and IPA…

Open weights apache-2.0
View model

Model · Text to speech

audio.cpp-gguf

Audio.cpp

This directory contains audio.cpp-native GGUF conversions of multiple speech models. These files are intended for use with audio.cpp. If you enjoy the project, please star audio.cpp on GitHub and this Hugging Face repository. For conversion details, supported layouts, direct-file loading, sidecar embedding, and the latest compatibility notes, see the audio.cpp GGUF guide: - https://github.com/0xShug0/audio.cpp/blob/main/docs/gguf.md!!! Converted and quantized packages are checked with automated metrics, but perceived quality can still differ for human listeners. Please validate the exact package, backend, and route to confirm the output is acceptable for your use case. The table lists the…

Open weights other audio.cpp
View model

Model · Text to speech

XTTS-v2

Coqui.ai

ⓍTTS is a Voice generation model that lets you clone voices into different languages by using just a quick 6-second audio clip. There is no need for an excessive amount of training data that spans countless hours. This is the same or similar model to what powers Coqui Studio and Coqui API. - Supports 17 languages. - Voice cloning with just a 6-second audio clip. - Emotion and style transfer by cloning. - Cross-language voice cloning. - Multi-lingual speech generation. - 24khz sampling rate. - 2 new languages; Hungarian and Korean - Architectural improvements for speaker conditioning. - Enables the use of multiple speaker references and interpolation between speakers. - Stability…

Open weights other coqui
View model

Model · Text to speech

chatterbox

Resemble AI

Chatterbox Multilingual V3 is the latest general-purpose multilingual TTS model in the Chatterbox family. It keeps the same 0.5B model size while improving speaker similarity, reducing hallucinations, and producing more natural, conversational speech across languages. V3 is designed for broad language coverage like V2, but with stronger stability and more expressive generation. It is the recommended multilingual model for users who want one voice cloning model that works across many languages. Try it in the Chatterbox Multilingual TTS V3 Space. Alongside V3, we are releasing the Single Language Pack: dedicated finetunes for priority languages where tighter quality control, stronger…

Open weights mit chatterbox
View model

Model · Text to speech

F5-TTS

Yushen CHEN

Download F5-TTS or E2 TTS and place under ckpts/ Paper: F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching

Open weights cc-by-nc-4.0 f5-tts
View model

Model · Text to speech

Kokoro-82M-v1.0-ONNX

ONNX Community

Kokoro is a frontier TTS model for its size of 82 million parameters (text in/audio out). First, install the kokoro-js library from NPM using: You can then generate speech as follows: Optionally, save the audio to a file: The model is resilient to quantization, enabling efficient high-quality speech synthesis at a fraction of the original model size.

Open weights apache-2.0 transformers.js
View model