How ScribeCue works

A small, independent tool for people who learn from video: students, researchers, writers and anyone who would rather read than scrub through a timeline.

Where YouTube transcripts come from

When you paste a link we read the caption track that YouTube already publishes for that video — either captions the creator uploaded or YouTube's automatic captions. We label which one you are looking at. We never download the video or its audio, and we never re-host it; the player on the page is YouTube's own embed.

If a video has no captions, there is nothing to read. In that case you can upload your own recording and we transcribe the audio.

Translation and AI summaries

Translations are written by an AI model from the transcript text, keeping every timestamp (free for videos up to 20 minutes). Summaries and chapters are written by a large language model from the transcript text only, with timestamps that link back to the moment in the video. Every summary point should be checkable against the transcript — click the time and listen.

Uploaded audio and video

Uploads are transcribed by a commercial speech-to-text engine with speaker separation. Files go straight from your browser to private storage, are deleted as soon as the transcript is ready (7 days at the latest), and are never used to train AI models. Minutes are counted from the file length, rounded up, and refunded automatically if a job fails.

What we will not do

• No video or audio downloader, ever.
• No reading of private, members-only or age-restricted videos.
• No account required for anything that is free.

Contact

Questions, corrections or a takedown request: hello@scribecue.com. See also privacy and terms.

Last updated 2026-10-18.