Remember what you watch.
Save any YouTube video. Rewatch reads the full transcript and turns it into linked notes you can search, study from and ask questions of.
- Free for up to 100 videos
- No card to start
- Works with any public video
Every video you save gets a one-line TL;DW
TL;DW · AI & ML
Layers of numbers, tuned until pixels become the right digit.
TL;DW · Education
Start with an empowerment promise and end with contributions, not “thank you”.
TL;DW · Programming
JavaScript runs one thing at a time; the event loop queues the rest.
TL;DW · Productivity
Deadlines wake the Panic Monster. Life’s biggest goals don’t have any.
TL;DW · Education
Connect the dots looking back, do what you love, remember time is short.
TL;DW · Programming
Build a GPT from an empty file and watch it learn Shakespeare.
TL;DW · AI & ML
Learning is nudging every weight a little downhill, over and over.
TL;DW · Programming
Python from first install to classes, in one sitting.
TL;DW · AI & ML
Layers of numbers, tuned until pixels become the right digit.
TL;DW · Education
Start with an empowerment promise and end with contributions, not “thank you”.
TL;DW · Programming
JavaScript runs one thing at a time; the event loop queues the rest.
TL;DW · Productivity
Deadlines wake the Panic Monster. Life’s biggest goals don’t have any.
TL;DW · Education
Connect the dots looking back, do what you love, remember time is short.
TL;DW · Programming
Build a GPT from an empty file and watch it learn Shakespeare.
TL;DW · AI & ML
Learning is nudging every weight a little downhill, over and over.
TL;DW · Programming
Python from first install to classes, in one sitting.
TL;DW · Programming
Python from first install to classes, in one sitting.
TL;DW · AI & ML
Learning is nudging every weight a little downhill, over and over.
TL;DW · Programming
Build a GPT from an empty file and watch it learn Shakespeare.
TL;DW · Education
Connect the dots looking back, do what you love, remember time is short.
TL;DW · Productivity
Deadlines wake the Panic Monster. Life’s biggest goals don’t have any.
TL;DW · Programming
JavaScript runs one thing at a time; the event loop queues the rest.
TL;DW · Education
Start with an empowerment promise and end with contributions, not “thank you”.
TL;DW · AI & ML
Layers of numbers, tuned until pixels become the right digit.
TL;DW · Programming
Python from first install to classes, in one sitting.
TL;DW · AI & ML
Learning is nudging every weight a little downhill, over and over.
TL;DW · Programming
Build a GPT from an empty file and watch it learn Shakespeare.
TL;DW · Education
Connect the dots looking back, do what you love, remember time is short.
TL;DW · Productivity
Deadlines wake the Panic Monster. Life’s biggest goals don’t have any.
TL;DW · Programming
JavaScript runs one thing at a time; the event loop queues the rest.
TL;DW · Education
Start with an empowerment promise and end with contributions, not “thank you”.
TL;DW · AI & ML
Layers of numbers, tuned until pixels become the right digit.
How it works
From a link to notes you’ll use
No transcripts to wade through. Rewatch reads every word and hands you the parts that matter, already connected.
1. Paste a link
Any YouTube video, Short or public playlist.
But what is a neural network?3Blue1Brown · 18:402. It reads the whole transcript
A TL;DW, summary, chapters, quotes and the key concepts, linked to your other videos.
TL;DW
Layers of numbers, tuned until pixels become the right digit.
- 01Each neuron holds a single number
- 02Weights decide what lights up next
Neural networkNeuronWeights and biases3. Find, link and ask
Search inside every transcript, follow ideas across videos, and ask your library anything.
See it work
Real videos, real notes
Pick a video. These are the kind of notes Rewatch writes, and every chapter opens YouTube at that moment.
TL;DW
Key points
- 01Each neuron holds a single number between 0 and 1, its activation
- 02Weights and biases decide how one layer’s activations drive the next
- 03This small digit recognizer has about 13,000 parameters to learn
- 04A whole layer is one matrix–vector multiplication plus a squashing function
Chapters
The reader
An hour-long talk, readable in a few minutes
Every video opens into a clean note, with the player and the full transcript alongside.
- TL;DW and summary. The one idea to remember, then everything worth knowing.
- Linked concepts. Every key idea links to the other videos that cover it.
- The full transcript. Follows along as the video plays. Click any line to jump there.
- Start a doc or add to chat. Turn a video into study notes, or ask about it.
Linked notes
Your videos, connected like a second brain
Rewatch turns the ideas in every video into linked notes, the way Obsidian does for writing. Open a concept and see what each video says about it, with the moment it comes up.
- Concept pages. People, tools and ideas, gathered from every video that mentions them.
- Backlinks and hubs. The ideas you keep coming back to rise to the top of the sidebar.
- Docs. Study notes that pull in videos, quotes and timestamps with @ and [[.
Features
Everything in one quiet place
Built for the videos you want to learn from, not the ones you scroll past.
Ask your library
ProOne chat for everything you’ve saved. Answers come from the full transcripts, with links to the exact moment. Add specific videos to focus it.
How do neural networks actually learn?
They measure how wrong they are with a cost function, then nudge every weight downhill Gradient descent7:20
Backpropagation works out which weights to blame, layer by layer Backprop, intuitively1:40
Mind maps
ProSee how a video’s summary, key points and topics connect.
Docs
Write study notes with videos built in: import a summary or chapters with @, link ideas with [[, and drop timestamps while you watch.
Exam prep: how networks learn
Core idea: a neural network is one big function, and learning is gradient descent on its cost.
Gradient descent, how neural networks learnWatch & write- Cost = how wrong the network is3:10
- Bias = how strong the input must be9:00
Graph view
ProEvery video, concept and doc as one map, connected by the ideas they share.
Search everything
⌘KJump to any video, concept or doc, or search inside every transcript and land on the moment it was said.
Videos
Gradient descent, how neural networks learn3Blue1BrownBut what is a neural network?3Blue1BrownInside summaries & chapters
Explain simply
One click rewrites a summary in plain language when a topic is new to you.
Playlist import
ProPaste a public playlist link and every video in it lands in your library at once.
Share summaries
ProTurn any video’s notes into a public page. Your own notes stay private.
Synthesize several videos
ProPick up to ten videos and get one brief: common themes, combined insights, what each adds and an order to watch them.
- But what is a neural network?
- Gradient descent
- Let’s build GPT
One brief
- Common theme: networks learn by adjusting weights
- Watch first: But what is a neural network?
- Then: gradient descent, then build GPT
- Each adds: intuition, the math, then real code
Who it’s for
For anyone who learns from YouTube
If you’ve ever thought “I watched a great video about this once”, Rewatch is for you.
Students
Turn lectures and explainers into notes you can revise from.
TL;DW
Learning means measuring how wrong you are, then stepping downhill.
Developers
Keep the one fix from a two-hour tutorial without rewatching it.
TL;DW
JavaScript runs one thing at a time; the event loop queues the rest.
Talks and podcasts
Get the ideas from long talks, with chapters to the best parts.
TL;DW
Deadlines wake the Panic Monster. Life’s biggest goals don’t have any.
Lifelong learners
Save how-tos and lessons with the steps pulled out as action items.
TL;DW
Start with an empowerment promise and end with contributions.
Organize
A library that sorts and links itself
Every video gets a category, tags and concepts automatically. Ideas that show up in more than one video become hubs, so related videos sit together without you filing anything.
- Gradient descent, how neural networks learnGradient descent2h ago
- What the heck is the event loop anyway?Event loopYesterday
- Inside the mind of a master procrastinatorProcrastinationMon
Pricing
Free to start. Pro when you’re hooked.
Save up to 100 videos for free. Go Pro for unlimited videos and the full toolkit.
Free
Everything you need to start a library.
No card required
- Save up to 100 videos
- Reads the full transcript of every video
- TL;DW, summary, key points and quotes
- Chapters that jump to the moment
- Linked concept notes across videos
- Study docs with timestamps and embeds
- Search inside every transcript
Pro
ProUnlimited videos and the full toolkit.
About $2.42/month, billed yearly
7 days free, then $29/year. Cancel anytime.
- Everything in Free, plus:
- Unlimited videos
- Chat with your whole library
- Graph view and mind maps
- Synthesize up to 10 videos
- Import public YouTube playlists
- Share links and library export
- Priority email support
- 7-day free trial
- Cancel anytime
- Secure payments by Stripe
What does Rewatch actually do?
You save YouTube videos to your library. Rewatch reads the full transcript of each one with Google’s Gemini AI and writes a TL;DW, a summary, key points, chapters, quotes and action items. The key ideas become linked concept notes, so videos that cover the same topic connect to each other, and you can search every transcript or ask your whole library a question.
Can I ask questions across all my videos?
Yes, with Pro. Open Chat in the sidebar and ask anything: Rewatch finds the relevant videos, reads their transcripts and answers with links to the exact moments. Use “Add to chat” on any video to focus the conversation on just those videos.
Does it work with Obsidian?
Rewatch works like Obsidian for videos: concepts are [[linked]] notes with backlinks and a graph. Docs, your study notes with timestamps and embedded videos, export to Obsidian-ready Markdown in one click.
Which videos work?
Public and unlisted YouTube videos, including Shorts and past live streams. Private videos and some age-restricted videos can’t be analyzed. Very long videos take a little longer.
How accurate are the notes?
They’re written by AI and are usually very good, but they can miss nuance or get details wrong. Every chapter links to that moment in the video, so you can always check the source.
What’s free, and what’s Pro?
Free includes up to 100 saved videos with full-transcript notes, linked concepts, Docs and search inside every transcript. Pro removes the limit and adds chat with your whole library, the graph view, mind maps, synthesis across videos, playlist import, share links and export.
How does the free trial work?
If you’ve never had Pro, both Pro plans start with a 7-day free trial. You add a card at checkout but aren’t charged until the trial ends, and you can cancel before then from Settings → Plan & billing.
Is my library private?
Yes. Your library and notes are only visible to you. A video becomes public only if you create a share link, and you can turn that link off at any time. You can delete your account and all your data from Settings.
Your next video deserves better than a forgotten tab.
Save your first video in under a minute. It’s free.
No card required · 100 videos free