Build YouTube chapter timestamps
Play the video, press M at each new topic, then copy the finished chapter list.
A manual marker that keeps you in the video
Pause or keep playing, then press M at every real topic change. Rename the markers in place and copy the finished block into the video description.
The four rules, checked as you go
YouTube only turns a description into chapters if the list satisfies all four: the first timestamp is 00:00, there are at least three chapters, the times ascend, and consecutive chapters are at least ten seconds apart. Break any one and the whole list silently stays plain text.
This tool checks all four while you work and tells you which one you are failing. That is the entire reason it exists — the rules are not hard, they are just invisible until you have published and the chapter bar is missing.
Why the first chapter must be 00:00
It is the rule people miss most. A list that starts at 00:42 because that is where the intro ends looks correct and produces no chapters at all.
The fix is to add a chapter at zero — call it Intro, or the title of the video — and put your real first topic second. The check flags this as soon as your earliest marker is later than the start.
Ten seconds is a floor, not a suggestion
Two chapters eight seconds apart invalidate the list. Not just that pair: the whole thing reverts to plain text, including the twenty chapters that were fine.
It bites on tutorials, where two short steps in a row are natural. When the tool flags a gap, either merge the two into one chapter or move the boundary — both are better than discovering it after upload.
Marked by hand, and that is the point
Press the key at each real topic change while the video plays, then name the markers. Nothing is generated: you decide where a chapter starts and what it is called.
Automatic chapter detection tends to find visual changes, which in a talking-head video is a camera cut and in a screencast is a window switch. Neither is a topic. The person who made the video knows where the topics are, and marking them takes about as long as watching it once.
Names are what viewers actually see
The chapter bar shows your text on hover, and the list appears under the video. "Setting up the database" is a reason to jump there; "Part 2" is not.
Write them as the answer to "what do I get if I click here". It is the same discipline as writing headings, and it is the difference between chapters that get used and chapters that get scrolled past.
The video stays on your machine
The file is played locally to mark it — it is never uploaded, and no account is needed. What leaves is a block of text you paste into your own description.
That matters for the normal workflow here: you are marking chapters before the video is public, which means marking a file that must not be handed to a third party yet.
- What makes a valid YouTube chapter list?
- Start at 00:00, include at least three chapters, keep timestamps ascending, and leave at least ten seconds between chapters. All four have to hold — the tool checks each one and says which is failing.
- Does this generate chapter titles with AI?
- No. You mark the real topic changes and name them yourself while watching the video. Automatic detection finds visual changes, and in a talking-head video that is a camera cut rather than a new topic.
- My chapters did not appear on YouTube. Why?
- Almost always the first timestamp is not 00:00, or two chapters are less than ten seconds apart. Either one silently turns the whole list back into plain text — including the chapters that were fine.
- How many chapters should a video have?
- Three is the minimum YouTube accepts. Past that, follow the video: one per real topic. Chapters every thirty seconds are noise, and a forty-minute video with three is not really chaptered.
- Where does the text go?
- Into the video description, as its own block of lines. Copy the finished list and paste it in; YouTube reads it from there and builds the chapter bar on the progress bar.
- Does the video upload?
- No. It plays locally so you can mark it, and nothing is sent anywhere. What you take away is a block of text.