Engine GPU (cuttlefish RTX 3060) OpenAI Whisper API Azure AI Speech (nb-NO)

API Key Used for this request only, never stored. Max 25 MB.

API Key Region

Model Fastest (small) Balanced (medium) Best quality ★ (large-v3)

Speakers Identify speakers Count

Vocabulary

Helps Whisper recognise technical terms. Not included in the transcript.

Drop audio file(s) here, or browse

MP3, WAV, OGG, M4A, FLAC, WEBM — max 200 MB per file

+ Add files

Advanced settings

Beam size 1 (fastest) 3 5 (best)

Controls search breadth — higher values improve accuracy but take longer. 5 is recommended for legal recordings.

VAD filter Remove silence

Voice Activity Detection — skips silent passages before transcribing. Speeds up processing and prevents the model hallucinating on silence.

Select a tool, run a request, and the result appears here.