About Scribe v2
Scribe v2 is a speech-to-text model that transcribes audio and video in more than 90 languages with word-level timestamps and speaker diarization.
It labels up to 32 speakers, tags non-speech events such as laughter or applause, and detects entities like names, numbers, and medical or financial terms with their timestamps. Keyterm prompting biases recognition toward product names, jargon, or proper nouns, and a no-verbatim mode drops filler words and false starts. It suits meeting and interview transcripts, subtitles, podcast and video indexing, and labeling speech datasets. Files can run up to 10 hours.
| Metric | Value |
|---|---|
| Max Input | 10 hours |
| Multilingual | Yes, 90+ languages |
Ready to build with Scribe v2?
Try Scribe v2 in the Workbench to prompt it, compare outputs, and iterate on prompts without writing any code. When you're ready to ship, call the same model from our API and build your own apps on top of it.
curl -sSf -X POST https://hub.oxen.ai/api/ai/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $OXEN_API_KEY" \
-d '{
"model": "elevenlabs-scribe-v2",
"messages": [
{
"role": "user",
"content": "Try sending a message."
}
]
}'Request parameters
Every field Scribe v2 accepts in the request body. See the API reference for response formats.
https://hub.oxen.ai/api/ai/chat/completions| Field | Type | Required | Default | Description |
|---|---|---|---|---|
input_audio | stringuri | Yes | — | Audio or video file to transcribe. Most audio and video formats are supported. |
language_code | stringnullable | No | — | Optional ISO 639-1 or 639-3 language code of the audio. Leave empty to detect it. |
diarize | boolean | No | false | Label which speaker is talking for each word. |
num_speakers | integernullable | No | — | Maximum number of speakers in the audio, if known. Helps diarization. Range: 1 – 32 |
timestamps_granularity | string | No | "word" | Timestamp detail in the transcript.nonewordcharacter |
tag_audio_events | boolean | No | true | Mark non-speech sounds such as (laughter) or (applause) in the transcript. |
no_verbatim | boolean | No | false | Remove filler words, false starts, and non-speech sounds from the transcript. |
keyterms | array<string>nullable | No | — | Words or phrases to recognize more accurately, such as product names or jargon. Up to 1,000 terms. |