AI · Web app
ElevenLabs Scribe
Speech-to-text API providing word-level timestamps and audio event tagging.
ElevenLabs Scribe is a speech-to-text model and API designed to transcribe complex and rapid audio with high accuracy. It returns detailed word- or character-level timestamps, speaker diarization, and non-verbal audio event tags. It is particularly useful for building interactive media experiences, lyric synchronizers, and fast-talking audio transcription tools.
What it's for
- Transcribe rapid or dense speech such as fast rap vocals and fast-paced podcasts
- Generate word-level timestamps for synchronized karaoke or lyric-highlighting interfaces
- Detect non-verbal audio events like laughter or footsteps alongside spoken words
- Separate and identify multiple speakers using diarization across audio files
Features
- Word-level and character-level timestamp granularity
- Automatic audio event tagging for sounds like laughter and applause
- Speaker diarization supporting up to 32 individual speakers
- Multichannel transcription with options for separate or combined time-aligned output
- Keyterms biasing to improve accuracy for technical terms or names
- Built-in entity detection and redaction for PII and sensitive data
Pricing
Freemium. Free tier available; usage-based API billing thereafter.
Details
- Type
- Web app
- Category
- AI
- Topic
- Speech Recognition
- Open source
- No
- Platforms
- API
- For
- Developers building transcription, captioning, or audio-aligned web applications.
- Alternative to
- OpenAI Whisper, Deepgram, AssemblyAI
- Tags
- stt, speech to text, transcription, audio, elevenlabs, api
Why it was saved
- @addyosmani: Shows that ElevenLabs Scribe (scribe_v1) accurately transcribes extremely fast speech (4.28 words/sec) with word-level timestamps and audio-event tagging, and provides a copyable pattern for building transcription apps.
Related
Fast local audio transcription benchmark and Python snippet for Apple Silicon Macs.
Open-source 600M English speech-to-text model with fast inference and timestamps
Ultra-low-latency streaming text-to-speech and instant voice cloning model.
Fast, low-cost text-to-speech API and studio for creators and developers
AI speech-to-text platform for multilingual transcription, translation, and video subtitling
Turn long-form audio and video recordings into written content assets.
Updated