One image that shows the whole video at once. Frames are taken at even intervals across the entire duration, laid out in a grid with the time on each, and written into a single PNG with a title bar. It is what you send instead of a link when someone needs to know what is in an hour of footage and has ninety seconds to find out.
One image, the whole video
The output is a single PNG: a header with your title on one side and the duration and frame count on the other, then the frames in a grid, each with the time it was taken from.
That format exists because of how it gets used. A link makes someone commit to watching; an image they can take in at a glance tells them what is there and lets them decide. It pastes into a message, an email or a document without anyone needing your tool or your file.
Even sampling, not detection
Frames are taken at equal intervals from the beginning of the video to the end. The tool does not look at what is on screen and does not try to find the interesting moments.
That is the right behaviour here even though it sounds worse. An even sample is an honest summary of the whole recording — you can see the pace, where the long static stretches are, whether anything happens after the halfway point. A tool that picked the interesting frames would hide exactly that.
Three settings that decide the shape
Frame count starts at 12, which fits most videos onto a sheet you can read without scrolling. Columns default to 4 for landscape footage and 3 for portrait, chosen from the source aspect so the sheet does not come out a strange shape. Cell width starts at 320 pixels.
Together they set the output size: cell width times columns, plus the header. Raise the cell width when it is going to print, lower it when it is going into a chat window, and raise the frame count when the video is long enough that twelve frames leave gaps you care about.
Timestamps make it an index
Each cell carries the moment it was taken from, in the corner. It costs almost nothing visually and changes what the sheet is: not just a picture of the video, but a map into it.
Someone looking at the sheet can say "what happens at 34:12" and go straight there. Without the times they can only say "the one with the chart", which is the difference between a reference and a decoration.
Vertical video gets a different grid
When the source is taller than it is wide, the column count drops to 3 automatically. Four columns of portrait cells would make a sheet far wider than it is tall, with everything too small to see.
It is a default rather than a rule — you can set the columns yourself. But it is the kind of thing that should not need thinking about, so the tool reads the aspect ratio and picks.
What people send it for
Showing a client what is in a batch of footage without uploading the footage. Giving a colleague the shape of a two-hour recording before they commit to watching it. Filing a visual record of a session next to the notes. Checking your own archive when the filename tells you nothing.
The common case is somebody who needs to make a decision about a video without watching it, which is most of the times a video gets sent to anybody.
PNG, and why not JPG here
The sheet is written as PNG because it has text on it. Titles and timestamps drawn over frames are exactly what the compression in JPG smears, and the sheet is a document to be read rather than a photograph to be looked at.
It is the opposite call from the frame-extraction tool, which writes JPG because those outputs are photographic frames with no text drawn on them. Same site, different format, and the reason is what is in the image.
Nothing leaves the machine
The video is read on your own computer, the frames are taken there, and the sheet is drawn in this tab before you download it. No upload, no account, nothing stored.
That is what makes it usable for the case it is best at: showing someone what is in footage that you are not allowed to send them.
- Can I choose how many frames appear?
- Yes. Frame count starts at 12, the column count defaults to 4 for landscape video and 3 for portrait, and each cell starts at 320 pixels wide. All three are adjustable before you generate.
- Is the storyboard watermarked?
- No. The PNG carries your title on the left of the header, the duration and frame count on the right, the frames, and the timestamps if you leave them on. Nothing else is added.
- Are the frames chosen or sampled?
- Sampled, evenly across the whole video from the first frame to the last. This tool does not look at what is on screen, which is why it gives an unbiased picture of the whole thing rather than a picture of the interesting parts.
- Then how is this different from the slide extractor?
- The extractor watches for the picture changing and then holding still, so it returns one image per slide however long each was shown. This returns twelve frames evenly spaced, regardless of content. For a lecture you want the extractor; for footage, this.
- How big is the output image?
- It follows your settings: the cell width times the column count, plus the header. Twelve frames at 320 wide in four columns is roughly 1300 pixels across. Raise the cell width for print, lower it for something to paste into a chat.
- Can I turn off the timestamps?
- Yes, there is a toggle and it starts on. Leaving them on is usually right, because a frame you want to see in motion is then a known point in the file rather than a search.
- Does it work on a vertical video?
- Yes. The column count switches to 3 automatically when the source is taller than it is wide, so the sheet stays a sensible shape instead of becoming a wide strip of narrow cells.
- Does the video upload?
- No. The frames are taken on your own machine and the sheet is drawn in this tab. Nothing reaches a server, and closing the tab is the deletion.
