Start with the source, not the export setting
A still image extracted from a video reflects the resolution, exposure and compression of the original recording. A 1920-pixel export cannot create fine details that a 720-pixel source never captured. Choose a clear source clip before adjusting export width. When a scene includes fast motion, select footage recorded with a short exposure time; otherwise edges will look blurry even if the resolution is high.
Choose a sensible number of frames
If you are producing a contact sheet or thumbnail candidates, extracting 12 to 24 frames from a short clip can give a useful overview. For a longer interview or presentation, a few evenly spaced stills may be sufficient. For exact action moments you may need a dedicated timeline editor. FrameKit intentionally uses even sampling: it does not claim to detect the aesthetically best moment.
Understand timestamps and video decoding
Compressed video formats often store some frames as differences from other frames. Seeking to a time means your browser may decode nearby data before it can display an image. As a result, extraction speed is not identical to video duration, and two browsers can handle the same input differently. MP4 with H.264 and WebM are widely supported; specialized phone codecs can vary by browser.
JPG or PNG?
JPG is practical for photos and video screenshots because it reduces download size. PNG is lossless and useful for flat graphics or diagrams, but files will usually be larger. The current FrameKit tool exports JPG images. If you need transparent images, first use a video editor or a dedicated background removal tool, since normal video clips rarely contain transparency.
Useful situations
Creators use extracted frames for thumbnails and shot references. Teachers can capture slides from recorded lessons. Small businesses can select stills from product clips for catalog drafts. For publishing images, check that you have the right to reuse the video content and any people or products shown.