Free YouTube Transcript Downloader

Three options. (1) Use the tool above: paste any public YouTube URL, click Get transcript, and download as TXT, SRT, VTT, JSON, or CSV in 1-3 seconds. Free preview — the first 200 words for longer videos, or the entire transcript for short videos under 200 words. Sign up free to unlock the full transcript on longer videos plus DOCX/PDF export. (2) Use YouTube's built-in transcript panel for one-off reading: click the three dots under the video, choose Show transcript, then copy-paste. Free, native, but no file formats and no batch. (3) Sign in to VexaScribe and upload the video file itself — useful when YouTube has no captions for the video, the auto-captions are wrong, or you need Whisper Large-v3 accuracy.

A few reasons. The creator may have disabled captions in YouTube Studio. YouTube's auto-caption system may not have run yet (it typically takes a few hours after upload, and longer for non-English languages). The video may be private, unlisted, age-restricted, members-only, or part of a YouTube Music or YouTube Kids context where captions aren't available. Live streams in progress don't have transcripts until they're archived. If the tool returns no captions, sign in and upload the video file to run full Whisper transcription.

YouTube uses an internal speech recognition model — not OpenAI's Whisper Large-v3. Typical accuracy on clear English audio sits around 85-92% Word Error Rate; it drops sharply on heavy accents, overlapping speakers, low-volume audio, technical jargon, or non-English where YouTube's auto-translate kicks in. Creator-uploaded human transcripts are generally accurate; auto-generated captions are not. If you're quoting verbatim or republishing under fair use, proofread first. For higher accuracy on the same audio, sign in and upload the video file — Whisper Large-v3 typically hits 95% on clean English.

On the URL paste tool above (Path A): no. You're limited to the caption tracks YouTube already has for that specific video. If a creator only uploaded English captions, you can't fetch French via this tool. YouTube's auto-translate feature is a player-side overlay, not a separate caption track. To get a different language: sign in and upload the video file — Whisper Large-v3 transcribes in 99 languages from the source audio, regardless of what captions exist on YouTube.

This isn't legal advice — talk to a lawyer for your specific case. Personal use (reading, studying, note-taking), fair-use quoting (short excerpts with attribution), and accessibility (captioning your own viewing) are generally fine in most jurisdictions. Republishing a full transcript as your own content without permission is a different question: YouTube's Terms of Service restrict reuse of platform content, and the underlying speech is the creator's copyrighted work. For commercial use, attribution-required quoting, or republishing, get the creator's permission first. The US Copyright Office's fair use guidance covers transformative use, but the line is fact-specific.

YouTube Shorts: yes if the Short has captions (auto or creator-uploaded). Many Shorts don't have captions because creators don't enable them. YouTube Music: typically no — music tracks don't carry caption tracks in the standard YouTube Music interface. Live streams: only after the stream ends and YouTube archives it. While the stream is live, captions may render on-screen but aren't downloadable via the standard caption API. Sign in to upload the recorded file for live-stream archives.

Not on the free URL paste tool above — it processes one URL at a time. For higher volume, sign in to VexaScribe and use the in-app bulk upload: drag and drop up to 50 audio or video files at once (MP4, MOV, MP3, M4A, WAV, FLAC, OGG, OPUS), choose your export formats, and download the batch as a single ZIP with original filenames preserved plus a CSV manifest. For YouTube specifically, download the videos first (or upload their audio tracks), then run bulk transcription. This is the right path when you're processing a back-catalog, an educator's course library, or doing accessibility retrofits across a whole channel.

SRT for video editors (Premiere Pro, DaVinci Resolve, Final Cut, CapCut) and for re-uploading captions to a different platform. VTT for HTML5 web players, styled captions, or HLS streaming. TXT for plain text — quoting in articles, feeding into an LLM, copy-pasting into Notion or Obsidian. JSON for developer pipelines and per-word timestamp precision (down to millisecond). CSV for spreadsheet-based content analysis and qualitative research workflows. If you're not sure: TXT for reading, SRT for editing video, JSON for building anything custom.

YouTube's built-in transcript panel is the simplest option for one-off reading: click the three dots under the video → Show transcript → copy-paste. Pros: zero tools, no rate limit, works on any device with the YouTube web app. Cons: copy-paste only (no SRT/VTT/JSON), no batch, no automatic language selection for multi-language videos, no per-word timestamps. Use the native panel when you just want to read; use our tool when you need file formats, multi-language download, video editor workflows, or no-existing-captions handling via Path B (signed-in upload). We don't hide that the native panel exists — for one transcript a day, it's a perfectly reasonable alternative.

On Path A (caption fetch), most tools in this category converge on similar quality because we're all fetching YouTube's existing captions — the underlying data is the same. UX, format selection, and language picker quality differ. Tools differentiate on Path B: real ASR on the video file itself when captions are missing or wrong. VexaScribe's Path B uses Whisper Large-v3, available via signed-in upload. DownSub, NoteGPT, NoteLM, Tubetranscript, and similar are honest free Path A tools; if you have a preferred UX, use it. We're aiming to be honest about Path A's limits and to offer a real Path B when the captions aren't there.