Audio Transcriber โ Podcasts, Calls & Meetings
Pricing
from $6.30 / 1,000 minute transcribeds
Turn recordings into text: podcasts, interviews, calls, meetings and voice notes in 90+ languages, with timecodes on every passage. Takes a file of any length or a podcast link. Speaker labels, SRT, RAG chunks. Export data, run via API, schedule runs, or integrate with AI workflows.
Paste a link from YouTube, TikTok, Instagram, Facebook, X, Vimeo, SoundCloud, Twitch, Dailymotion, VK, Rutube or Bilibili โ the audio track is resolved for you โ or a direct file: MP3, M4A, WAV, FLAC, OGG, MP4, MOV, WEBM and anything else ffmpeg can read. Video is stripped to its audio track automatically.
Upload one audio or video file from your computer instead of giving a URL.
Fast is the everyday choice. Accurate makes fewer mistakes on accents, noise and technical vocabulary; it is slower, and the price per minute is the same.
Two-letter code of the language spoken, for example en, ru, id. Leave empty to detect it automatically.
Return an English transcript no matter what language is spoken.
Mark who is speaking in each part of the recording. Every segment gets a speaker ("Speaker 1", "Speaker 2", โฆ), the transcript is laid out as a conversation, and subtitles carry the names. Charged per minute on top of the transcription, and slow: the speaker model runs on the machine, so expect several minutes of processing per hour of audio, while a plain transcript is near-instant.
How many people talk in the recording. Leave empty and the Actor works it out on its own; set it when you already know, for example two people in an interview.
Names, jargon or spellings that appear in the audio, comma separated. Helps the engine get proper nouns right.
Return every recognised passage with its start time and duration, alongside the full transcript.
Split the transcript into retrieval chunks that keep the timecode where each one starts.
Target size of one chunk. Chunks break on sentence ends, so real sizes vary.
How many seconds of the previous chunk to repeat at the start of the next one.
Also return a ready-made subtitle file as text.
Optional. Supply your own key and the per-minute price drops sharply; you then pay Groq directly for the recognition. Leave empty to use the built-in key with nothing to set up.
How many files to transcribe at the same time.
Files larger than this are skipped with an error row instead of being downloaded.
