Skip to content
SPECIFICTION
Get listed

MetaVoice-1B

themetavoice.xyz

WebLinuxAPISelf-hosted
Visit themetavoice.xyz
Spec score 7.7 #144 of 223 in Text-to-Speech Software

Spec sheet

Purpose
MetaVoice-1B is a 1.2-billion-parameter text-to-speech model trained on 100,000 hours of speech.Read 13 Oct 2026 · github.com
Platforms
web,api,self_hostedRead 20 Sep 2026 · github.com
Deployment
The project provides a web UI and inference server that can be deployed on AWS, GCP, or Azure.Read 12 Oct 2026 · github.com
Free plan
YesRead 20 Sep 2026 · github.com
API
The inference server provides API definitions at its /docs route.Read 13 Oct 2026 · github.com
Integrations
The README describes an optional Weights & Biases integration for logging fine-tuning runs.Read 13 Oct 2026 · github.com
Support
The repository directs users to its active GitHub issues for contributions and discussion.Read 13 Oct 2026 · github.com
License
The model is released under the Apache 2.0 license, which the project describes as usable without restrictions.Read 13 Oct 2026 · github.com
Access options
The page lists local use, Hugging Face, and a Google Colab demo as ways to use the model.Read 12 Oct 2026 · github.com
API access
YesRead 20 Sep 2026 · github.com
Commercial use
YesRead 20 Sep 2026 · github.com
Dataset caveat
The fine-tuning workflow does not check for overlap between training and validation datasets, so the page advises keeping them disjoint.Read 12 Oct 2026 · github.com
Expressive speech
The project prioritizes emotional speech rhythm and tone in English.Read 12 Oct 2026 · github.com
Fine-tuning
The project supports fine-tuning the first-stage language model and describes cross-lingual voice cloning with fine-tuning.Read 13 Oct 2026 · github.com
Fine-tuning data limit
The fine-tuning instructions say the project does not check for overlap between training and validation datasets.Read 13 Oct 2026 · github.com
Hardware requirement
Installation requires a GPU with at least 12 GB of VRAM, Python 3.10 to below 3.12, and pipx.Read 12 Oct 2026 · github.com
Hardware requirements
Installation prerequisites include a GPU with at least 12 GB of VRAM and Python 3.10 or later but earlier than 3.12.Read 13 Oct 2026 · github.com
Performance
After compilation, the synthesise API runs faster than real time on Ampere, Ada-Lovelace, and Hopper GPUs, according to the project page.Read 12 Oct 2026 · github.com
Quantization limit
Experimental int4 and int8 inference can run faster, but the project says this degrades audio quality.Read 13 Oct 2026 · github.com
Speech style
The model is designed for emotional speech rhythm and tone in English.Read 13 Oct 2026 · github.com
Startup time
The reference inference script takes 30 to 90 seconds to start, depending on hardware, because it compiles the model.Read 12 Oct 2026 · github.com
Tracking integration
The project offers optional Weights & Biases logging, enabled through configuration and installing the relevant dependencies.Read 12 Oct 2026 · github.com
Voice cloning
It supports zero-shot cloning of American and British voices using 30 seconds of reference audio.Read 13 Oct 2026 · github.com
Ways to use
The project documents a local reference implementation, a web UI, an inference server, Hugging Face usage, and a Google Colab demo.Read 13 Oct 2026 · github.com

Where Specifiction places it

Alternatives to MetaVoice-1BAll alternatives →