About this tool
What is the Audio Transcriber?
Upload an audio file (MP3, WAV, M4A, OGG, FLAC, up to 25 MB) and the Whisper-powered transcription returns the full text with sentence-level timestamps. Accurate enough for podcasts, interviews, voice memos, meeting recordings, and voice-over scratch tracks.
Free, rate-limited at 10 transcriptions per hour. Files are processed via Groq's Whisper Large v3 API and immediately discarded.
FAQ · AUDIO TRANSCRIBER
Questions about the Audio Transcriber
How accurate is the transcription?
Whisper Large v3 is the state-of-the-art for general-purpose transcription — typically 95%+ word accuracy for clear speech in supported languages. Accuracy drops with heavy accents, overlapping speakers, or low-quality audio.
What's the file size limit?
25 MB per file (roughly 30 minutes of MP3 at standard quality, or longer with compressed bitrates). Longer files: split first or upgrade to a paid Whisper plan for higher limits.
Is my audio stored or used for training?
No. Audio is sent to Groq for transcription and dropped immediately. Groq has a no-training-on-API-inputs policy. We don't store the audio on ReelStack's servers.