YouTube · TikTok · Instagram · your own files
AI Video Summarizer
Paste a link and read the video instead of watching it — key points, chapters and the full transcript, with every claim linked back to the second it was said.
Try one
Watch less, understand more
A 45-minute talk takes 45 minutes to watch and about 90 seconds to read. Paste the link and you get the argument, the chapter structure and the full transcript at once — enough to decide in seconds whether the video deserves the other 44 minutes.
Links, audio files and your own uploads all go through the same pipeline. A conference talk, a podcast episode, a recorded call and a handful of links stop being four different tools with four different outputs.
45 min → 90 sec
Every point keeps a link to where it came from
A summary is only worth anything if you can check it. Every point here carries the second it was derived from: click it and the video jumps there, hover it and the matching line lights up in the transcript. Verifying is one click, not a second viewing.
That is the real difference between summarizing here and pasting a transcript into a chat assistant. The assistant gives you fluent text with no route back to the source — so when it drifts, there is no way to notice. Here the source is attached to every line, and the transcript sits beside the summary the whole time.
It matters most exactly where people skip the checking: technical talks, market commentary, medical and legal content. Any plausible-sounding paraphrase can quietly invert what was actually said.
…so people can focus on creativity, strategy and empathy.
02:19 · transcript
Three depths, ready before you ask
Gist is a few sentences — enough to decide whether this is worth your time. Brief gives you the argument and its main supporting moves. Full notes is close to what a careful person would have written while watching.
Move between the three freely, with no wait, because all three are prepared together. Most people start at Gist and go deeper only where it turns out to matter — which is a different way of working from choosing a summary length before you know what is in the video.
Paste once, take any format
From the same result you can take a timestamped transcript, a chapter list, a plain-text summary, SRT or VTT subtitles, study flashcards, a mind map, or the comment section as a CSV — without sending the video through a second time.
Output language is set separately from what is spoken, so a Japanese talk can produce an English summary while the transcript stays in Japanese. Nothing is lost in translation because nothing is replaced by it.
YouTube, TikTok, Instagram, and your own files
Paste a YouTube link, a TikTok, an Instagram Reel, or drop in an mp4, mov, mp3 or wav from your own machine. All of them come back the same way: a timestamped transcript, a summary at three depths, chapters, and whatever else you need from the same result.
Short clips and long recordings are handled the same way. A forty-second Reel and a three-hour podcast both come back in seconds, and the transcript is clickable line by line in either case.
What people use it for
Students
Turn recorded lectures into notes you can revise from, then into flashcards for the week before the exam.
Researchers and analysts
Scan conference talks and earnings calls for the parts worth your attention, with a citable timestamp on every claim.
Content and social teams
Break down what a competitor's video actually says, and turn your own back catalogue into posts, captions and transcripts.
Anyone with a backlog
Get through a watch-later list the way you get through a reading list — by reading it.
Questions people ask before they paste
Is this AI video summarizer free?
Reading is. Every video already on the site opens in full, with no account. Processing a new one spends credits, and a new account starts with ten to try it on. A video whose captions already exist costs a flat rate; one we have to transcribe from the audio is priced per ten minutes.
Do I need an account?
Not to read. Anything already processed — the summary, the transcript, the timestamps — opens for anyone. You need an account to put a new video through, because that is the part that costs us a provider call.
What can it summarize?
YouTube, TikTok and Instagram links, and your own video or audio files up to 100 MB. Anything with speech in it works; the whole pipeline is built around spoken content rather than text documents.
How is this different from asking ChatGPT to summarize a video?
The timestamps survive. A chat assistant returns fluent prose with no route back to the source, so when it drifts you have no way to notice. Here every point carries the second it came from, and the full transcript stays on screen beside the summary.
Does it work on TikTok and Instagram?
Yes, and that is the part most tools skip. Neither platform publishes a transcript to anyone, so there are no captions to fetch — we run the audio through speech recognition ourselves and give you the words with real timings.
Can I get the summary in a different language from the video?
Yes. Output language and spoken language are independent, so a Spanish video can produce an English summary, or an English one a Japanese summary, while the transcript stays in the original.
How long can a video be?
There is no length limit. A three-hour podcast comes back in roughly the time a ten-minute one does. Length changes the price only when we have to listen to the audio, where it is charged per ten minutes.
Does it work on private or age-restricted videos?
No. We can only read what is publicly reachable. If you can play something yourself but we cannot get to it, download it and upload the file instead — that path has no such restriction.
Is the transcript included?
Always. The full timestamped transcript comes with every summary, it is searchable, and you can export it on its own as plain text, Markdown, SRT or VTT.
Start with one video
Ten credits when you sign up — enough to work through a handful of videos properly. No card required.
Summarize a video