ScribeGrab

ScribeGrab turns audio and video into text: a transcript, SRT/VTT subtitle files, or an MP4 with the captions burned into the picture. Around 100 languages, auto-detected. Files up to 2 GB and 90 minutes — MP3, WAV, M4A, FLAC, OGG, MP4, MOV, MKV, WEBM and more. No account, no watermark, no per-minute charge and no daily cap. It runs on our own GPUs. Clear speech transcribes very accurately; heavy accents, crosstalk, background noise or poor recordings are harder for any tool, so give it a proofread. When it hears more than one voice it splits the transcript into speaker turns automatically, up to eight speakers — those labels are estimated by AI and can slip on crosstalk. It can also produce a censored version with swear words beeped, muted or cut, and you see the full word list with timestamps before anything is changed. Files are deleted the moment they are done and results are wiped within 45 minutes. An optional euro 5 month pass adds meeting notes, real speaker names and Word export; the free tool runs the same model and does not expire.