What it is
A GitHub repository for MetaVoice-1B, an AI model trained for text-to-speech with a focus on emotional speech and the ability to clone American and British voices with as little as 1 minute of training data.
Gabriel’s notes
MetaVoice-1B is a 1.2B parameter base model trained on 100K hours of speech for TTS (text-to-speech). It has been built with the following priorities: Emotional speech rhythm and tone in English. No hallucinations. Zero-shot cloning for American & British voices, with 30s reference audio. Support for (cross-lingual) voice cloning with finetuning. We have had success with as little as 1 minute training data for Indian speakers. Support for long-form synthesis. We’re releasing MetaVoice-1B under the Apache 2.0 license, it can be used without restrictions. Includes a DEMO!
Good fit if you want to:
- build, test, or ship software faster (APIs, dev tooling, code assistance).
- learn a new skill, concept, or workflow with structured guidance.
Pricing snapshot (auto-enriched): Open-source model available for free under Apache 2.0 license; usage-based pricing applies for hosted services such as narrating 100,000 words for $4.29; no mention of per seat pricing or free tier for hosted services.
Work-use / compliance snapshot (auto-enriched): MetaVoice-1B is an open-source TTS model released under Apache 2.0 license suitable for use without restrictions, but there is no explicit information regarding workplace data handling, training usage, retention policies, SSO availability, or compliance with SOC2, HIPAA, or GDPR.
Alternatives (auto-enriched): Alternative: ElevenLabs | Comparison: ElevenLabs offers advanced AI voice synthesis with a focus on naturalness and emotional expression, similar to MetaVoice but with a broader commercial application focus.
Before you adopt it: check the README, license, recent commits, and open issues to gauge maintenance and fit.
Author: metavoiceio
Note: pricing and policy details can change—verify on the official site before making decisions.