Extract Frames from Video
Pull still images out of a video automatically. Take a frame every few seconds, spread a set number evenly across the video, or let scene detection grab one frame for every new shot or slide.
- Files stay on your device
- No sign-up
- Free to use
Which frames
Higher sensitivity catches smaller changes; lower values keep only clear cuts.
Images
Processing happens on your device with FFmpeg. Keep this tab open; long or high-resolution videos can take several minutes.
How to use Extract Frames from Video
- Drop a video onto the upload area.
- Choose how frames are picked: at an interval, a fixed number, or at scene changes.
- Set the interval, number of frames or detection sensitivity, and the image width.
- Select Extract frames and download the images as a ZIP.
Extract Frames from Video features
Interval extraction
One frame every second, few seconds or minute.
Evenly spaced overview
A fixed number of frames spread across the whole video, perfect for contact sheets.
Scene change detection
Captures a frame whenever the picture changes significantly, ideal for slide decks and edited films.
Adjustable sensitivity
Tune scene detection from clear cuts only to subtle changes.
Resized JPGs
Choose the original size or smaller images for faster browsing.
In your browser
FFmpeg runs locally; the video is not uploaded.
When to use Extract Frames from Video
- Saving every slide from a recorded presentation or lecture.
- Creating a storyboard or shot list from an edited video.
- Making a visual summary of long surveillance or dashcam footage.
- Choosing a thumbnail from evenly spaced candidates.
- Building image datasets from video for training or analysis.
Extract Frames from Video FAQ
How does scene change detection work?
FFmpeg compares each frame with the previous one and computes a score from 0 to 1 for how different they are. Frames with a score above the threshold set by the sensitivity slider are saved, along with the very first frame.
Why did scene detection miss a change?
Gradual transitions such as fades and slow pans change the picture a little at a time, so no single frame scores highly. Raise the sensitivity, or use interval extraction for such footage.
How many frames can I extract?
Up to 200 per run. For long videos with frequent changes, lower the sensitivity or use a longer interval.
What is the difference from Video to JPG?
Video to JPG lets you pick moments by hand. This tool selects frames automatically across the whole video.
Can I get PNG frames?
Use Video to PNG to export a lossless image sequence from a chosen range.
Turning hours of video into a handful of images
Video holds far more frames than anyone needs to review: an hour at 30 frames per second is 108,000 images. Sampling reduces that to something useful. Fixed intervals are predictable and work for steady footage such as timelapses, security cameras or long recordings. A fixed count spread evenly gives an overview of any video regardless of length.
Scene detection is smarter for edited content. Films, tutorials and slide presentations consist of shots or slides that each last a while. Measuring how much each frame differs from the previous one and saving a frame only when the difference is large captures exactly one image per shot or slide, so a 45-minute lecture might produce thirty images, one for every slide.
The difference score combines changes in brightness and colour across the frame. Hard cuts produce scores near 1, while camera motion and people moving within a scene produce smaller values. The sensitivity slider sets the threshold between them; there is no single correct value, so try a couple of settings for your footage.
Extracted frames are JPG images sized for browsing. For archival or editing work at full quality, choose the original width, or use Video to PNG for a lossless sequence.