Universal 3 5 Pro
Quick start
# Inspect the price — a plain request returns the 402 challenge:
curl -i https://api.glianalabs.com/v1/infer \
-H "content-type: application/json" \
-d '{
"model": "assemblyai/universal-3.5-pro",
"audio_url": "https://example.com/input.mp3"
}'
# Pay + run in one step with the mppx CLI (create a wallet: npx mppx account create):
npx mppx https://api.glianalabs.com/v1/infer \
-J '{"model": "assemblyai/universal-3.5-pro", "audio_url": "https://example.com/input.mp3"}'Parameters
The URL of the audio or video file to transcribe, or a data URI. Pass a public URL; to use a local file, upload it (free) via POST /v1/media (≤40MB) and pass the returned url.
Timestamp in milliseconds at which to stop transcription.
Timestamp in milliseconds at which to begin transcription.
Whether to extract key phrases.
Whether to detect sensitive content.
Content safety confidence threshold from 25 to 100.
Whether to include filler words such as um and uh.
Domain-specific model for medical terminology.
Whether to detect named entities.
Whether to filter profanity.
Whether to apply text formatting.
Whether to classify topics using IAB categories.
Words or phrases to prioritize during transcription.
Language code of the audio.
Whether to automatically detect the language.
Whether to transcribe each audio channel separately.
Natural-language instructions for transcription style.
Whether to add punctuation.
Whether to redact personally identifiable information.
Whether to generate an audio file with spoken PII redacted.
Format of the redacted audio file.
PII categories to redact.
Whether to include unredacted fields alongside the redacted transcript.
Replacement strategy for redacted PII.
User-defined terms to redact, grouped by label.
Whether to analyze sentence sentiment.
Whether to identify speakers in the transcript.
Expected number of speakers.
Minimum fraction of speech required for transcription.
Controls transcription randomness from 0 to 1.