Built because the obvious shortcut did not work

Why ChapterGen exists

I wanted chapters for a long video before publishing it. My first try was the obvious one: upload the file to ChatGPT and Claude and ask for timestamps.

It did not work. The video file was too large to upload, the assistants could not actually listen to the audio, and when I pasted partial text the timestamps that came back were invented. Getting usable chapters still meant transcribing the video, splitting it by topic and checking every timestamp by hand.

So I built a tool that does exactly that, end to end, from the file itself.

How it works

ChapterGen extracts the audio from your file and transcribes it with OpenAI's Whisper speech-recognition model, so every timestamp comes from the real audio. A language model then finds where the topic changes and writes a descriptive title for each section.

The output follows YouTube's chapter rules: it starts at 0:00, has at least three chapters, and no chapter is shorter than 10 seconds. You can edit every title before you copy it.

What makes it different

Most chapter tools need a public YouTube URL. ChapterGen works from the file, so you can add chapters to drafts, private and unlisted videos, and audio-only podcasts before anything is published.

It supports 90+ languages, files up to 500 MB, and you can try it on 2 videos for free with just your email.

Your files

Uploaded files are deleted from our servers after processing. We keep the transcript and the chapters so you can come back to them. The privacy policy explains the details.