Extract frames from a video as JPG or PNG images
Need stills from a clip — reference shots for a storyboard, images for a training dataset, the frames around a goal, or every slide from a recorded presentation? This free video frame extractor turns an MP4, MOV, MKV or WebM video into a set of JPG, PNG or WebP images and packs them into a single ZIP file.
Pick how frames are chosen: one every few seconds, every Nth frame, every single frame, at a list of exact times, or automatically at each scene change. You see how many images you will get before you start, and every file name carries its frame number and timestamp. Decoding happens in your browser, so the video is never uploaded.
How to use the Video Frame Extractor
- Drop a video on the box. Its duration, resolution and codecs are read immediately.
- Choose “Which frames”: every N seconds, every Nth frame, every frame, at times you type, or at scene changes.
- Optionally limit the range with “From” and “To”, and choose the image format, quality and width.
- Check the plan — it shows how many frames you will get and roughly how much space they take — then press “Extract frames”.
- Download everything as a ZIP, or click any thumbnail to save that frame alone.
Video Frame Extractor features
Five ways to pick frames
Interval in seconds, every Nth frame, every frame, a typed list of timestamps, or automatic scene-change detection.
Exact frames, not screenshots
Frames are decoded with WebCodecs through Mediabunny (MPL-2.0), so you get the real frame at each time at full resolution, with phone rotation applied.
Scene detection
Consecutive frames are compared with colour histograms; when the difference passes your sensitivity threshold, that frame starts a new scene.
Clear count before you start
The plan shows the number of frames and an estimated size, and runs are capped at 1,000 images to keep your browser responsive.
Timestamp file names
Files are named like clip_0042_00-01-23.500.jpg, so they sort in order and you always know where each came from.
Private
Nothing is uploaded; the ZIP is built in your browser with JSZip (MIT licence).
Choosing the right mode
| Mode | Good for | Example |
|---|---|---|
| Every N seconds | Overviews, datasets, time-lapse stills | 1 frame every 2 s from a 10-minute video = 300 images |
| Every Nth frame | Animation and motion study | Every 6th frame of 30 fps video = 5 images per second |
| Every frame | Short clips, rotoscoping, frame-by-frame analysis | 3 s of 30 fps video = 90 images |
| At times | Specific moments you already know | 0:12, 1:05.5, 3:40 |
| Scene changes | Slides from recorded talks, shots of an edited film | One image for each cut |
Format, quality and size
JPG at 85–92% looks identical to the video frame and is small. PNG is lossless — use it when you will edit the frames or need exact pixels, but expect files of several megabytes each at 1080p. WebP is smaller than JPG at similar quality and works in all modern browsers and apps.
Remember that video frames are already compressed, so saving as PNG does not add detail that the video does not have. Reducing the width to 1280 or 640 px makes extraction faster and the ZIP much smaller — ideal for previews and machine-learning datasets. For a single perfect still, the Video Thumbnail Generator can also suggest the sharpest frame; for one overview image, use the Video Contact Sheet.
Video Frame Extractor FAQ
How do I convert a video to JPG images?
Drop the video here, choose “Every N seconds” (or another mode), keep JPG as the format and press “Extract frames”. Download the ZIP to get all the images.
How do I extract every frame of a video?
Choose “Every frame”. The plan tells you how many images that is — about 30 per second for most phone videos — and up to 1,000 frames can be extracted per run. Use “From” and “To” to work through a longer video in parts.
How does scene-change detection work?
Each frame is shrunk and its colour histogram compared with the previous frame’s. A big difference means a cut. Lower the sensitivity percentage to catch softer transitions; raise it to keep only hard cuts.
Are the frames full quality?
Yes, at “Original” width each image has the video’s full resolution. PNG keeps the decoded pixels exactly; JPG and WebP add a little compression you can control with the quality slider.
Why is the frame count in the plan an estimate?
Many phone videos have a variable frame rate. The plan estimates the count from the frame rate measured on the frame timestamps; “Every frame” and “Every Nth frame” then save exactly the frames that are in the file.
Is there a size limit?
Videos are read in pieces, so file size is not the issue; the 1,000-frame cap per run keeps memory use reasonable on phones.