About yt-transcript

What we do

yt-transcript turns any public YouTube video into clean, timestamped text. Paste a link and you get the transcript — from the video’s existing captions when they exist, or from on-device AI speech recognition (Whisper-class) when they do not. You can copy it, download it as TXT, SRT, VTT, or CSV, translate it into 100+ languages, batch up to 25 videos at once, or feed it to a model through our REST API or MCP server.

It is a text pipeline, not a video editor. No project, no timeline, no media import — URL in, text out, in the format you need.

How it works

Captions when they exist

For videos that already have a caption track — human-written or auto-generated — we read the captions directly from YouTube’s own data. No Google Data API key, no OAuth, no quota management on your side. One Authorization: Bearer header and you are in. This is the fast, cheap path: 1 credit per video, no matter its length.

AI transcription when they do not

When a video has no captions, we download its audio and run Whisper-class speech recognition on our own GPU hardware. It is asynchronous — the call returns a job ID, you poll until it is done. This path costs more (1 credit per 90 seconds of audio) because it runs real compute, and it handles 99+ languages.

Everything runs on our own hardware

Transcription runs on our own GPUs in the EU. We do not call a third-party transcription API. Your uploaded media is deleted as soon as transcription finishes — only the resulting text is kept. Your content is never used to train AI models.

What we do not do

• No ads. No trackers. No analytics of any kind.
• No selling or renting of your data.
• No third-party data sharing except Stripe (payments) and our hosting provider (operational infrastructure).
• No AI training on your content or transcripts.

Who it is for

• Researchers & students — quote and cite exactly; every line carries its timestamp, and batch mode covers an entire reading list of talks.
• Content creators — repurpose video into articles, or generate bilingual subtitles for one video in two languages.
• Developers & pipelines — the REST API fetches transcripts at scale (25 videos per request) for indexing, moderation, or RAG corpora.
• Language learners — follow along with translated SRTs — original line above, translation below — while the video plays.

What you get

• Timestamped transcripts in TXT, SRT, VTT, CSV, or JSON.
• AI summaries: TL;DR, key points, takeaway.
• Translation into 100+ languages, including bilingual SRT.
• Batch mode: up to 25 videos per job, playlist expansion.
• Study tools: quizzes, flashcards, mind maps, study guides.
• REST API with Bearer keys + an MCP server for AI agents.
• A free Chrome extension — one-click transcripts right in YouTube.

Get started

25 free transcripts without any account. A free account tops that up to 50. No credit card, no trial clock. Try it now →