whereismyai
MOSS Transcribe
MOSS Transcribe is an open-source audio understanding model for multi-speaker transcription, diarization, and timestamps in 50+ languages.
MOSS-Transcribe-Diarize is an open-source audio understanding model designed to turn real-world long-form audio into structured, speaker-aware transcripts in a single pass. It combines a Qwen3-0.6B causal decoder with a Whisper-Medium encoder to jointly perform speech transcription, speaker diarization, and timestamp generation. The model supports over 50 languages and handles long, multi-speaker recordings like meetings, calls, podcasts, interviews, lectures, and video content. It accepts raw audio inputs and outputs compact timestamped transcripts with speaker labels. Users can deploy the model using Python, SGLang Omni, vLLM, or the FunASR ecosystem, and utilize a local subtitle web app for review and export.
MOSS Transcribe at a glance
Key features
- Long-form multi-speaker transcription
- Speaker diarization with consistent speaker labels
- Precise timestamp generation in seconds
- Acoustic event awareness
- Custom prompt and hotwords support
- Local subtitle web app for upload and review
- OpenAI-compatible endpoints via SGLang Omni and vLLM
Use it for
- Transcribing multi-speaker meetings and calls
- Generating speaker-aware podcast and interview transcripts
- Processing long-form lecture and video recordings
- Exporting subtitle files like SRT, ASS, and JSON
Alternatives to MOSS Transcribe
- MOSS-TTS — Delivers zero-shot voice cloning and fine-grained pronunciation control.
- WhisperX — It enables 70x faster, batched transcription with high-accuracy alignment.
- Whisper Diarization — Combines OpenAI Whisper with voice activity detection and speaker embedding.
- WhisperLive — It provides real-time speech-to-text using multiple backends like Faster Whisper and TensorRT.
- Maybloc — Separates up to 25 speakers and preserves overlapping speech.
- faster-whisper — Transcribes audio faster with lower memory usage than standard implementations.
- AssemblyAI — Industry-leading speech models deliver high accuracy and low latency for voice applications.
- Whisper ASR Webservice — Provides a REST API for multiple ASR engines including WhisperX and Faster Whisper.
- Whisper Timestamped — Generates accurate word-level timestamps without extra neural networks.
- FireRedAudio — A unified model handling understanding and generation with decoupled continuous representations.
- transcribe.cpp — Runs 16 model families and 60+ variants with GPU and CPU acceleration.
- Moonshine — Runs entirely in your browser with no audio leaving your device.
All MOSS Transcribe alternatives