A free video transcript generator for YouTube, TikTok, Instagram and a thousand other sites. Export timecoded SRT and VTT, or plain text — and grab the video file itself while you're here.
Turn speech into TXT, SRT, or VTT — uses existing captions when available.
Every part of the chain is exposed — the probe, the format list, the caption source, the export. Nothing is hidden behind a single “magic” button.
If the video already ships captions — creator-uploaded or auto-generated — those exact lines are used, so the wording matches the source and the job finishes fast. Only when there are none does Whisper transcribe the audio.
Two paths, one result
Every segment carries a start and an end, so exports drop straight onto an editing timeline or load as a caption track. Plain TXT is there when you just want the words.
TXT · SRT · VTT
yt-dlp does the extraction, so support is not a short hand-picked list. YouTube, Instagram, TikTok, X, Reddit, Vimeo, Facebook, Twitch and far beyond.
1000+ extractors
Fetch the format list and choose exactly what you want: a resolution preset, an MP4 / WebM / MKV container, audio-only extraction, or a specific yt-dlp format ID when the preset isn't precise enough.
Up to 1080p
No signup wall, no email capture, nothing stamped onto your output.
Paste-in cookies unlock YouTube HD, live only for that one job, and are deleted the moment it finishes.
Copy the URL of any public video — a YouTube watch link, an Instagram reel, a TikTok post — and paste it into the input. No account, no extension, no upload.
Transcribe reads the words out of the audio. Download saves the video file itself, with resolution and container options once you fetch the format list.
Take the transcript as plain TXT, or as timecoded SRT or VTT subtitles ready to drop straight onto a timeline. Video downloads save as MP4, WebM, or MKV.
All three come from the same transcript, so you can export one, then another, without re-running the job. The only real difference is whether each line carries a timecode.
| Format | Timecodes | Best for | Opens in |
|---|---|---|---|
| .txt | No | Reading, searching, quoting, pasting into docs or an LLM | Any text editor |
| .srt | Yes | Editing timelines, burning in subtitles, uploading captions | Premiere, Resolve, Final Cut, YouTube |
| .vtt | Yes | Web players and HTML5 caption tracks | Browsers, video.js, HLS players |
Extraction runs through yt-dlp, so the list is measured in thousands rather than a handful of logos. If the link is public, it will almost certainly resolve.
Including YouTube, Instagram, TikTok, X / Twitter, Reddit, Vimeo, Facebook, Twitch, Dailymotion, SoundCloud, LinkedIn, and Bluesky — plus roughly a thousand more sites covered by yt-dlp extractors.
Straight answers about accuracy, formats, supported sites, and what happens to the cookies you paste in.
Paste the video's public link into the input above and press Transcribe. The engine pulls the audio, reads the speech, and returns a timecoded transcript you can export as TXT, SRT, or VTT. Most short clips finish in well under a minute.
Yes. Transcribing a video and exporting TXT, SRT, or VTT is free, with no account and no watermark on the output. There is no per-minute quota to track and nothing is charged for downloading the transcript files.
Anything yt-dlp can read, which covers well over a thousand sites — YouTube, Instagram, TikTok, X, Reddit, Vimeo, Facebook, Twitch, and Dailymotion among them. The link has to be publicly reachable; private and login-walled posts will not resolve.
When a video already ships captions, those exact captions are used, so accuracy matches whatever the creator or the platform published. Without captions, the audio is transcribed by Whisper, which is strong on clear speech and weaker on heavy accents, crosstalk, and loud music.
TXT is the words alone, best for reading, search, or pasting into a document. SRT and VTT are subtitle formats that pair each line with a start and end timecode, so they can be imported into an editor or loaded as a caption track on a player.
YouTube serves its higher-resolution streams only to signed-in sessions. Without a session, the best available stream is roughly 360p. Pasting your own exported cookies authenticates the request and unlocks 720p and 1080p for that download.
Yes. Switch to the Download tab, fetch the format list, then choose a resolution and a container — MP4, WebM, or MKV. Audio-only extraction is available too, and advanced users can select an exact format ID from the full list.
Cookies are used only for the download you paste them into, held for the life of that job, and deleted when it finishes. They are never shared. Media files are written to job storage for retrieval and are not published anywhere public.
One input, two outputs: a transcript you can search and cut from, or the video file itself. No account, no install, nothing to cancel later.