speech-to-text (ASR) and speaker diarization
Open-source voice AI suite centered on VibeVoice-ASR, a multilingual speech-to-text model that transcribes long-form audio with speaker diarization and timestamps.
Solves the same problem with a different architecture or at a different layer. Expect to rewrite the integration.
same tasks in a Swift/CoreML library for Apple platforms, not a Python drop-in
Frontier CoreML audio models in your apps — text-to-speech, speech-to-text, voice activity detection, and speaker diarization. In Swift, powered by SOTA open source.
use it for a full training/inference framework to customize ASR/TTS models, heavier than a pretrained model
A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
use it for a modular ASR toolkit with streaming, VAD, punctuation, diarization and training, not one unified pretrained model
Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
use it for an Apple Silicon MLX/CoreML voice toolkit with ASR/TTS/diarization; different language/runtime
AI speech toolkit for Apple Silicon — ASR, TTS, speech-to-speech, VAD, and diarization powered by MLX and CoreML
use it for low-latency streaming STT with VAD/wake-word, a slim library around Whisper rather than a unified long-form model
A robust, efficient, low-latency speech-to-text library with advanced voice activity detection, wake word activation and instant transcription.
An earlier or more limited way to do the job. Still the right call when you need something small, proven, or CPU-only.
use it for lightweight ASR on Mandarin/Cantonese/EN/JA/KR with emotion/event detection; fewer languages and no long-form diarization
Open-source SenseVoiceSmall model for Mandarin, Cantonese, English, Japanese, and Korean ASR, language ID, emotion recognition, and audio event detection.
use it for extremely small multi-task audio (ASR/TTS/VC) on edge; capacity/accuracy below frontier ASR
[AutoArk] GPA (General Purpose Audio) can do ASR, TTS and voice conversion with one tiny model!
a simultaneous speech recognition/translation/synthesis model; narrower than VibeVoice for long-form diarized ASR
StreamSpeech is an “All in One” seamless model for offline and simultaneous speech recognition, speech translation and speech synthesis.
a lightweight Rust reimplementation of a mini real-time ASR/TTS model for native/browser edge, more limited than VibeVoice
Voxtral ASR & TTS running natively and in the browser. A Rust implementation of Mistral's Voxtral mini realtime ASR / TTS using the Burn ML framework
Does the same job, but ships as an app. Useful to a person, not swappable into a codebase.
an end-user voice studio app for cloning/dictating, not a library/model to integrate
The open-source AI voice studio. Clone, dictate, create.
a local end-user voice app for cloning/dubbing/transcription, not embeddable
VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.
a voice productivity assistant app, not a speech-to-text model/library
🎙️ An intelligent voice productivity assistant that turns speech into clean text, useful actions, and structured knowledge. It helps users capture ideas, communicate naturally, automate repetitive tasks, and stay productive across different apps and workflows.