Video to Structured Notes
Turns a long video into notes with timestamps, using Gemini's native video input rather than a transcript.
What it is for
Working from a transcript throws away everything that was on screen. This uses native video input to produce notes with timestamps, so a diagram or a demo is in the notes rather than a line saying someone pointed at something.
The skill
Watch the video and produce notes, not a transcript. Format: ## Summary Three sentences: what the video is about, who it is for, what it concludes. ## Timeline `[mm:ss]` — one line per section, naming what happens, not what is said. ## Claims worth checking Each claim, its timestamp, and whether the video supports it with evidence shown on screen or only asserts it. ## What was demonstrated Anything performed on screen — commands, settings, steps — written out so it can be followed without rewatching. Include exact values shown. Rules: - Timestamps must come from the video, not be estimated from the runtime. - If the audio is unclear at a point that matters, write `[unclear]` with the timestamp instead of guessing the words. - Skip filler: introductions, sponsor reads, subscribe requests.
How to install it
Gemini 1. Open Gems → New Gem. 2. Paste the text below into the instructions field. 3. Save. The Gem appears in the sidebar and keeps the instructions across conversations. In Google AI Studio the same text goes in System instructions.