VibeVoice
MIT-licensed long-form multi-speaker text-to-speech.
About
Microsoft's open TTS model (1.5B) that can synthesize up to 90 minutes of continuous multi-speaker speech — up to four distinct voices — from text, with cross-lingual narration and even basic singing. It drew 54k+ GitHub stars and runs on ~7GB VRAM, so a single consumer GPU handles it.
One honest caveat: Microsoft later removed the official repository after misuse, so the project now lives on through a community fork under the same MIT license.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: 54k+ GitHub stars; MIT license confirmed on the repo; 1.5B TTS; up to 90 minutes of multi-speaker speech; ~7GB VRAM for inference; Upstream removal after misuse is the key risk — maintenance now depends on the community fork; English and Chinese training data only. License: MIT — permissive. Note: Microsoft removed the official repo; the community fork continues under the same MIT license..
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 9.0/10
- Maintenance · recency, releases, issue load
- 5.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · 54k+ GitHub stars; MIT license confirmed on the repo
- · 1.5B TTS; up to 90 minutes of multi-speaker speech; ~7GB VRAM for inference
- · Upstream removal after misuse is the key risk — maintenance now depends on the community fork
- · English and Chinese training data only