Transcribe Audio to Text
Create a draft transcript with fal-ai/whisper, automatic language detection, word-level timing, and TXT/SRT/VTT export. Review names, accents, noise, and overlapping speech.
AI transcription runs on secure servers.
Whisper-powered speech-to-text creates word-level timestamps and TXT, SRT, and VTT downloads. It costs 1 credit per minute. A free account includes 6 AI audio credits per month with no card required. FileMorf disables provider request-payload storage, keeps provider-hosted output private to its provider account, requests one-hour provider expiry, and keeps the encrypted job input and FileMorf result copy until the expiry shown in Recent Results.
Need more? A 1-day pass starts at $3, with no subscription. See plans
How it works
Add your audio
Drop in an MP3, voice memo, interview, or any recording. The spoken language is detected for you.
Transcribe with Whisper
Whisper AI writes out the words with timestamps down to the word — 1 credit per minute of audio.
Export TXT, SRT, or VTT
Grab plain text for reading or timed captions for video. Workspace access to encrypted retained server objects expires on your plan’s schedule. Expired objects are queued for automatic cleanup, and deletion is retried if storage removal fails.
Frequently asked questions
How accurate is the transcription?
FileMorf uses fal-ai/whisper with automatic language detection and word-level timing. Accuracy varies with language, names, accents, noise, music, and overlapping speakers. Treat the transcript as a draft and review it before quoting, captioning, or publishing.
What formats can I export?
Download plain text (TXT) for a clean read, or timed captions as SRT and VTT for video subtitles. Word-level timestamps are captured too, so captions line up tightly with the audio instead of drifting across long lines.
Is my audio private?
Transcription sends the rendered edit over TLS for a background server job. FileMorf disables provider request-payload storage and does not use your audio to train models. FileMorf asks fal.ai to expire its private generated-media copy after one hour and disables request-payload storage; access to FileMorf’s encrypted job input and result copy follows the workspace schedule (7 days on free accounts and no later than 30 days on paid plans), followed by retried automatic cleanup.
How much does it cost?
Transcription costs 1 credit per minute of audio. A free account includes AI audio credits every month — enough to transcribe a short recording, no card required — and Pro plans add a larger monthly allowance.
Cleaner audio transcribes better — try AI noise removal or remove silence first, or open the full Audio Studio.