microsoft/VibeVoice

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

54432 stars · 6132 forks observed · 2026-09-21

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-21, confidence 0.97

VibeVoice is an open‑source framework for high‑quality text‑to‑speech and speech‑to‑text generation. It combines autoregressive and diffusion models to produce expressive, long‑form, multi‑speaker audio.

Use cases

• generate podcast‑style audio with multiple speakers from a script
• create long‑form narrated audiobooks with natural turn‑taking
• perform real‑time streaming speech transcription for meetings
• run speech‑to‑text inference on edge CPUs without a GPU
• synthesize multilingual dialogue with expressive emotions

When to choose

• You need high‑fidelity, expressive TTS for long passages or multi‑speaker content
• You want a unified solution that supports both TTS and streaming ASR
• You prefer a Python library that can run on CPU‑only environments
• You are building applications that require low‑latency, continuous transcription

When to avoid

• You require a lightweight TTS model for very short utterances only
• You need a fully managed cloud service with built‑in scaling
• Your deployment target is non‑Python environments without a binding layer

Facets

library · maturity active

tts llm-inference audio-processing artificial-intelligence speech-processing python windows cloud multi-speaker long-form diffusion continuous-tokenizer edge-cpu streaming-asr speech-processing natural-language-processing linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="microsoft/VibeVoice")