Scribe v2 RealtimePricing
$0.39 per hour of input audio, billed per second, rounded up to the next whole second.
Input Modalities
- Audio
Output Modalities
- Text
Frequently asked questions
More models from ElevenLabs
See all ElevenLabs models →ElevenLabs Dubbing v1 translates and dubs audio or video into target languages while preserving the original speakers' voices. Call it via the ElevenLabs-compatible project API /v1/dubbing/project; billed by source media duration per target language.
ElevenLabs Dubbing v2, the higher-quality generation of dubbing, translates and dubs audio or video into target languages while preserving the original speakers' voices. Call it via the ElevenLabs-compatible project API /v1/dubbing/project; billed by source media duration per target language.
ElevenLabs Voice Isolator: removes background noise, music and ambience to extract clean speech. Uses the native ElevenLabs endpoint POST /v1/audio-isolation; input audio must be at least 4.6 seconds; billed by input duration.
English-only voice changer (speech-to-speech): converts input audio into a target voice while keeping the original intonation and emotion. Uses the native ElevenLabs endpoint POST /v1/speech-to-speech/{voice_id}; official SDKs work by changing base_url. Billed by input duration.
Ultra-fast English-only speech synthesis optimized for real-time use, with ~135ms latency and up to 30,000 characters per request. Call it through the OpenAI-compatible /v1/audio/speech endpoint with an ElevenLabs voice_id as voice; outputs mp3, opus, wav or pcm, with chunked streaming via stream_format=audio.
ElevenLabs' fast, affordable speech synthesis model, with ~100ms latency across 32 languages and up to 40,000 characters per request, at half the per-character price. Built for voice agents, interactive apps and bulk conversion. Call it through the OpenAI-compatible /v1/audio/speech endpoint with an ElevenLabs voice_id as voice; outputs mp3, opus, wav or pcm, with chunked streaming via stream_format=audio.
