# FastScribe

> FastScribe at fastscribe.org transcribes uploaded audio/video and supported public social-video links. It is independent of other services using the FastScribe name.

## Capabilities and limits

- Upload audio or video files up to 1 GB per file. A paid account does not raise this upload limit.
- Public URL inputs: YouTube, TikTok, Instagram, X/Twitter and Facebook. Private, restricted or unavailable media may fail even when viewable in a signed-in browser.
- Vimeo and SoundCloud require a file you already have permission to use; their page URLs are not supported. Bilibili is not supported.
- Export TXT, DOCX, XLSX and SRT. Review transcript wording and subtitle timing before publication.
- Choose the spoken language or automatic detection. Language options include regional variants; support is not an accuracy guarantee for every dialect.
- Speaker labels, punctuation and timestamps depend on the audio and processing result. Noisy audio, overlapping speakers and language switching can reduce accuracy.
- Free credits are available. See current pricing for allowances; uploaded files and social URLs consume credits at different rates.

## Entry points

- [Audio to Text](https://fastscribe.org/audio-to-text)
- [Video to Text](https://fastscribe.org/video-to-text)
- [MP3 to Text](https://fastscribe.org/mp3-to-text)
- [Facebook to Text](https://fastscribe.org/facebook-to-text)
- [Spanish transcription and sample](https://fastscribe.org/es/speech-to-text/spanish)
- [Arabic transcription and sample](https://fastscribe.org/speech-to-text/arabic)
- [Language directory](https://fastscribe.org/speech-to-text)
- [Pricing](https://fastscribe.org/pricing)
- [Machine-readable pricing](https://fastscribe.org/pricing.txt)

## Sample evidence

The Spanish and Modern Standard Arabic samples are original scripts spoken by synthetic voices and processed on 2026-09-07. Their audio, reference text and unedited recognition output are available on the language pages. They test clean single-speaker audio only, not natural interviews or dialect accuracy. The Arabic result contains errors, including an omitted negation, and one long subtitle cue requiring review.