Grab a frame
Pull one still out of a video at full resolution, without screenshotting your own screen.
Your file
Drop a file here, or click to choose
It stays on this device. Nothing is sent anywhere.
Settings
Better than a screenshot
Screenshotting a playing video captures whatever your screen shows: scaled down, possibly with player controls over it, and with your display's color handling baked in. Extracting the frame gives you the actual pixels at the actual resolution.
Set the time to the exact moment you want. Scrub the video in any player, note the timestamp, and type it in as m:ss or h:mm:ss.s.
Which format
PNG is lossless and the right pick for anything you will edit or crop further. JPEG is much smaller and fine for a thumbnail. WebP splits the difference and is the smallest of the three at matching quality.
Landing on the right frame
The time field takes m:ss or h:mm:ss, and decimals work, so 1:23.4 is valid and useful. That precision matters more than it sounds: at 30 frames a second there are thirty distinct pictures inside every second, and on anything moving they are not interchangeable. A face mid-blink and the same face a tenth of a second later are very different thumbnails.
The reliable method is to scrub in any player, note the timestamp, then adjust by tenths from there. If the moment is hard to pin down, export an image sequence across a short range instead and pick from the results, which is faster than guessing six times.
Why the frame looks different from your player
Occasionally the extracted frame is a shade darker or flatter than what you saw. That is usually colour range: video is often stored with a limited range, and players expand it for display while a raw frame export does not. It is not damage, and it is why matching a thumbnail exactly to a paused screen sometimes takes an adjustment with brightness and contrast.
The other difference is size. A screenshot captures your display, scaled to fit a window, possibly with player controls across it. This reads the actual encoded pixels, so a 4K video yields a 4K still whatever size the window was.
Making a thumbnail people will click
Leave the width at 0 to keep the source resolution, then work from there. YouTube wants 1280 by 720, most platforms crop link previews to 1200 by 630, and both will be seen at a fraction of that size in a feed. Choose a frame that reads at thumbnail scale: a clear subject, strong contrast, nothing important near the edges where it may be cropped.
If the shape is wrong, crop the video first and then grab the frame, or grab the frame and reshape it with resize image. Take PNG if any editing follows, since each JPEG save adds artefacts and a thumbnail usually gets edited at least once.
Common questions
I got a black frame. Why?
The time you asked for is probably before the first frame or past the end. Try a second or two later.
Can I export several frames at once?
Yes, use the image sequence tool, which exports frames on an interval and hands them back as a zip.