Extract frames from video

Save any moment of a video as a JPG or PNG at the video's full resolution, the exact frame and not a blurry screenshot, or take one every few seconds and download them as a ZIP. It all happens on your device, so your video is never uploaded.

A free tool from Clyper, the AI clipping tool that turns long videos into captioned vertical clips.

Get started

Opening the video…

0:00.000 / 0:00.000

Keys: ← → one frame, Shift + ← → one second, space to play or pause. Click the bar to jump.

MP4, MOV or WebM, or drop one on the player. Frames are decoded on your device: nothing is uploaded.

An example: 8 seconds of IMPAULSIVE

Save this frame

Quality92

JPG at 90 or more looks the same as PNG at a fraction of the size.

Extract many frames

One every

Open a video to choose.

How to extract frames from a video

Open your video, or drop it on the player. MP4, MOV and WebM work, from a phone, a camera or a screen recording. The example shows how it works before you choose one.
Find the moment. Drag the bar, or play and pause. Then step one frame at a time with the arrow keys (or the buttons), and a second at a time with Shift.
Save the frame. Pick JPG (with its quality) or PNG and press Save frame. It's named after the video and the moment, like interview-01m23.500s.jpg.
Or take many. Choose one frame every so many seconds or frames, set where to start and stop, and press Extract. Click a thumbnail to save that one, or download them all as a ZIP. Up to 300 frames an extraction.

Phone videos filmed upright are often stored sideways with a note to turn them, and the tool reads that note, so the stills come out the right way up. For more ways, on a Mac, on Windows and on a phone, see how to get a still image from a video, in full quality.

How to get a still from a video in high quality (and why screenshots look worse)

Take the frame from the file, not from the screen. A frame grabbed from the video has exactly as many pixels as the video: a 1080p video gives a 1920 × 1080 image, a 4K one 3840 × 2160. A screenshot only has the pixels the player took up on your screen, and usually more problems besides:

A screenshot of a video player compared with a frame taken from the video file
	Screenshot of the player	Frame from the file (this tool)
Size	The player's size on screen, often smaller than the video	The video's full resolution
Which frame	Whatever was showing when you pressed the keys	The exact frame you stepped to
What's in it	Can catch play buttons, captions, a cursor or the progress bar	Only the picture
Source	A streaming player may be showing a lower-quality copy	The file itself, decoded once
Saved as	Your system's screenshot format, sometimes compressed again	JPG at the quality you pick, or lossless PNG

Two things no tool can fix: a frame can't be sharper than the video, and a fast-moving moment is blurred in the video itself. Step a few frames either way. In talking videos, the frame just as a word ends, with the mouth closed and the eyes open, is usually the sharpest and the most flattering.

The size of a still from common video resolutions
Video	Still from the file	Megapixels
720p	1280 × 720	0.9 MP
1080p (Full HD)	1920 × 1080	2.1 MP
1440p	2560 × 1440	3.7 MP
4K (2160p)	3840 × 2160	8.3 MP
Vertical 1080p (TikTok, Reels, Shorts)	1080 × 1920	2.1 MP

Video to JPG or PNG: which to choose

JPG for sharing and thumbnails; PNG for editing. Both keep the frame's full size. They differ in how they store it:

JPG and PNG for frames from a video
	JPG	PNG
Compression	Lossy: drops detail the eye barely sees, more at lower quality	Lossless: every pixel as decoded
File size	Small; usually several times smaller than PNG at quality 90	Large, especially for 4K or grainy footage
Best for	Thumbnails, covers, posts, contact sheets, sending	Editing and saving again many times, print, frame-by-frame analysis
Uploads	Every app and site takes it	Most apps take it; some limit the file size

A video is already compressed, so a PNG doesn't add detail; it just doesn't lose any more. That's why the tool starts on JPG at 92. YouTube's thumbnail upload from the app is limited to 2 MB, which a 1080p JPG fits easily.

Extracting every frame, or one every few seconds

A video has far more frames than it looks: a minute at 30 frames a second is 1,800 images. Picking a step keeps it useful. Here's what each gives from one minute of 30 fps video:

Frames from one minute of 30 fps video, by step
Step	Frames from a minute
Every frame	1,800
Every 15 frames	120
Every 30 frames	60
Every 0.5 seconds	121
Every 1 second	61
Every 5 seconds	13
Every 10 seconds	7

Each extraction takes up to 300 frames, so it stays quick and fits in your browser's memory; for more, run it again on the next range. Every frame is for studying motion or picking the one perfect frame of a fast moment. Every few seconds is for a contact sheet, a storyboard, or finding which part of a long video is which. Steps in frames are exact; steps in seconds take the frame on screen at each moment.

Use a frame as a thumbnail or cover

A frame of your own video is the most honest thumbnail there is, and the easiest: a clear face, mid-reaction, is what most good thumbnails are built on. Save it as JPG, then:

For a 16:9 YouTube thumbnail, open it in the YouTube thumbnail maker and add a few bold words. More in how to make a YouTube thumbnail.
For a 9:16 cover on Shorts, TikTok and Reels, use the Shorts thumbnail maker, and see how to change a YouTube Shorts thumbnail.

Stills show you what a video looks like. To find the moments in it worth posting, what's said matters more: Clyper reads a long video's transcript, picks the moments that stand on their own and makes each one a vertical clip, captioned and ready to post.

Questions

How do I extract a frame from a video?

Open the video in the tool above, move to the moment with the bar, then step frame by frame with the arrow keys or buttons until it's the one you want. Pick JPG or PNG and press Save frame. You get the exact frame at the video's full resolution, made on your device.

How do I get a high-quality still from a video?

Take the frame from the video file rather than a screenshot of a player. A screenshot is only as sharp as the player on your screen, and it can catch overlays and a streaming copy at lower quality. Saving the decoded frame keeps every pixel the file has: 1920 × 1080 for a 1080p video, 3840 × 2160 for 4K.

Why does my screenshot of a video look blurry?

Because it captures the screen, not the video. If the player is smaller than the video, the screenshot has fewer pixels; if it was moving, you may catch a blurred in-between frame; and players and streaming sites often show a lower-quality copy to save data. Grabbing the frame from the file avoids all three.

Should I save frames as JPG or PNG?

JPG for almost everything: at quality 90 or more it looks the same as PNG and the file is usually several times smaller. Choose PNG when you'll edit the image a lot and save it again, or need every pixel exactly as decoded. Neither adds detail the video didn't have.

How do I extract all frames from a video?

Choose Frames, then "every frame", set the range and press Extract. Each extraction makes up to 300 images, which is 10 seconds of 30 fps video, so take a long video in several ranges or pick a bigger step, like one frame a second.

Is my video uploaded?

No. Your browser reads the file and decodes the frames on your device with WebCodecs, the video decoder built into the browser. Nothing is sent to Clyper or anyone else.

Can I use a frame as a YouTube thumbnail?

Yes, and it's a good start: YouTube suggests a 16:9 image, which a frame from a 16:9 video already is. Save it as JPG, then add a few bold words in the YouTube thumbnail maker. For Shorts, TikTok and Reels covers, use the Shorts thumbnail maker.

Find the moments worth clipping

Paste a YouTube link and Clyper finds a long video's best moments and makes up to 30 vertical clips, cut on whole sentences, framed on whoever's talking and captioned.

See plans and pricing