Voice
Speech-to-text (STT)
Speech-to-text transcribes spoken audio into written text. Paired with a chat model and TTS it forms a voice conversation loop.
Language detection, file size caps and pricing per request or per minute are the practical parameters.
On eroq
Scribe One: 1 credit per started minute of audio, files up to 8 MB, automatic language detection.
Questions
What does transcription cost?
1 credit per started minute of audio on Scribe One — a voice note is 1 credit, an hour of audio 60.
What audio can Scribe One take?
One wav or mp3 file up to 8 MB — roughly eight minutes of compressed speech — with the language detected automatically or hinted with a code. Longer recordings have to be split; billing is 1 credit per started minute of each file, so splitting only costs the rounding.
Does transcription return timestamps or speaker labels?
No — the response is the text of the audio and nothing else. There is no diarization, no word-level timing and no subtitle file; captions get built from the text downstream.