How to Extract Frames From a Video and Save Them as Images

You can extract frames from a video entirely in the browser: add a clip, pick a moment or a set of moments, choose PNG, JPG or WebP, and download the stills. The video is decoded by the browser and drawn to a canvas, so it is never uploaded and no plugin is needed.

Unlike a GIF converter, this tool needs no transcoding engine. It decodes the clip with the browser's own video support and copies the picture to a canvas, so it starts instantly instead of downloading roughly 30 MB of WebAssembly first. That makes it the right tool for pulling a thumbnail, a slide image or an article illustration out of a clip. You can take one exact moment, one frame every N seconds, or a fixed number of frames spread evenly across the video.

Pulling a still image out of a video sounds like a job for a video editor, but the browser already knows how to play the clip, so it can also freeze it. The Extract Frames tool uses that ability and nothing else. It moves the video to a moment you choose, copies that picture to a canvas, and hands you an ordinary image file. There is no engine to install and no upload to wait for.

Extract Frames

Save a still from any clip as PNG, JPG or WebP. The browser decodes the video, nothing is uploaded and no engine is downloaded.

Open the tool

What people extract frames for

Most video edits never need the video itself. A thumbnail needs a single picture. A written tutorial needs a still of the button the reader is about to press. A bug report needs the exact second the error appeared on screen. In each case the clip is only the source, and the file you want at the end is an image that opens anywhere.

A manual screenshot can work, but it copies whatever is on the screen: the player window size, the controls, the browser chrome and any notification that happened to appear. Extracting the frame pulls the picture straight from the file, at the video's own resolution, with no interface around it. That is the difference between a still you can drop into a document and a photo of a video player.

Thumbnails and cover images

One frame is usually all a thumbnail needs, and single moment extraction is built for it. Scrub the preview to a moment that reads clearly at small size, and save it as PNG or JPG. Because the frame comes from the file, it is sharp at the full resolution of the clip rather than at the size of your window.

Step images for guides and tutorials

A guide often needs the same screen at several points, numbered in order. Fixed count mode spreads a chosen number of frames evenly across the clip, which is the quickest way to get that set without hunting for each timestamp. One frame every N seconds works better when the steps are unevenly spaced and you would rather sample the whole recording.

Illustrations for articles and posts

An article about a film, a game or a product launch needs a picture that shows the thing being discussed. A still taken from the source avoids the re-compressed look of a screenshot grabbed from a streaming page, and WebP output keeps the file small enough to load fast on a phone.

The three ways to grab frames from a video

The Extract Frames tool offers three sampling modes, and they answer different questions. Pick the one that matches how well you know the moment you want.

Single frame at one time

Type a time in seconds, down to a hundredth of a second, or drag the slider and the preview follows along so you can see the picture before you save it. This is the mode for a thumbnail or one specific shot, and it is the only mode that is not affected by the batch limit.

One frame every N seconds

Set an interval, one second by default, and the tool walks the clip from start to finish, saving a still at each step. Use it to watch how a scene changes, to build a contact sheet, or when you do not know exactly where the moment you want sits in the timeline.

A fixed number of frames, spread evenly

Ask for ten frames and the tool spaces them evenly from the first moment to the last. This is the fastest route to a storyboard strip or a small set of samples, and the count is adjustable up to thirty in one run.

How to extract frames from a video without uploading it

The whole flow runs on your machine and takes three steps.

  1. Add your video. Drop an MP4, WebM, MOV or MKV file onto the drop zone, or click to browse. The clip is opened from your disk and read by the page, and the browser reports the duration and resolution. Nothing is sent anywhere.
  2. Choose when to grab frames. Pick Single frame, Every N seconds or Fixed count, then set the time, the interval or the number. Choose PNG, JPG or WebP, and optionally a width. For JPG and WebP a quality slider appears, and the frame counter shows how many stills the current settings will produce.
  3. Extract and download. Press Extract frames and watch the progress as each picture is decoded in turn. When it finishes, download each frame on its own, or press Download all, which saves the set with a short pause between files so the browser does not block the batch.

Extraction is free and unlimited, and the source video is never altered, shortened or re-encoded. Only the images you export are new files.

Output format, quality and size

Three choices shape what you get back, and none of them touch the original clip.

FormatBest forQuality slider
PNGLossless stills, text, interfaces, further editingNo, always lossless
JPGSmall files that open anywhereYes, 50 to 100, default 92
WebPSmallest size for the webYes, 50 to 100, default 92

Separately, you can scale the frame down to 1920, 1280, 854 or 640 pixels wide, and the height follows to keep the aspect ratio. The default is the original size. A source narrower than the width you choose is kept at its own size rather than enlarged, so a small clip is never stretched into a blurry still. Scaling keeps the whole picture, just smaller, so it is the shortest path to a frame that fits a slide or a web layout.

Frames, GIF or a screen capture

When the goal is to show motion, a still is the wrong output. Three options cover most cases, and they are not interchangeable.

ApproachOutputBest forWatch out for
Extract FramesPNG, JPG or WebP stillsThumbnails, step images, article picturesA still cannot show movement
Video to GIFAn animated GIFA short silent loop that plays anywhereNo audio, 256 colours, grows fast with length
Screen captureA picture of the screenA quick note when the file is not handyCopies the player window and controls too

The rule is simple. If the reader needs to see one moment, extract the frame. If they need to see the motion, make an animated GIF from a short clip instead, because a still can never carry movement. And if you only need to point something out in a chat, a screen capture is fast enough, as long as you accept everything the player was showing.

Video to GIF

When you need motion instead of a still, turn a short clip into a clean animated GIF. Also runs in the browser.

Open the converter

What this tool cannot do

These are the edges of the tool, and they matter when you are deciding whether it fits the job.

  • Frame timing follows the video's own frame positions. Compressed video stores keyframes and encodes the frames between them, so the picture you get can differ from the exact timestamp you asked for by a few milliseconds. The tool takes the time to a hundredth of a second and shows the frame in the preview, but it cannot invent a frame the video does not hold.
  • Thirty frames per run in the batch modes. Fixed count and Every N seconds stop at thirty images. Single frame mode is not affected. If you need more, download one batch and run the tool again on another section of the clip.
  • Long videos are limited by your device memory. Every extracted frame is held as a full image until you download it, so sampling a very long clip at a small interval can make the tab heavy. Keep the batch small on long files.
  • No cropping inside the frame. The size options change the overall width, not the framing. To keep only part of a picture, extract the full still and crop it afterwards with the image cropper.
  • The clip must be decodable by the browser. MP4 with H.264 or H.265, WebM and most MOV files work. A file the browser cannot decode is reported as a failed frame rather than saved as a broken image.
  • It never enlarges a frame. Choosing a width larger than the source keeps the source size, so this is not an upscaler. If a clip is low resolution, the still will be too.

FAQ

Open the Extract Frames tool, drop in your clip and pick one of the three modes. Single frame takes one moment you type or scrub to, Every N seconds walks the clip and saves a still at each step, and Fixed count spreads a chosen number of frames evenly across the whole video. Then choose PNG, JPG or WebP, optionally pick a width, and press Extract frames. Each saved frame appears as its own card with a Download button, and Download all saves the whole set with a short pause between files. The video is read from your device and decoded by the browser, so nothing is uploaded. There is no account, no watermark and no cost per run. The first run needs no engine download either, because extraction uses the browser's own decoder rather than a transcoding library.
Use PNG when you want the frame exactly as the video holds it, because PNG is lossless and keeps every pixel. It is the right choice for text, interfaces, sharp edges and anything you plan to edit further. Use JPG when you need a small file that opens anywhere, and lower the quality slider to shrink it more. Use WebP when the image is going on a website and you want the smallest size at good quality. The quality slider appears for JPG and WebP only, from 50 to 100, and defaults to 92. PNG ignores it entirely. As a rule, a still of a screen recording or a document should be PNG, while a photographic frame from a camera clip can be JPG or WebP without a visible difference. You can convert between formats later with the image converter if you change your mind.
Because compressed video is stored as keyframes plus the differences between them, and a browser can only show a picture it has fully decoded. When you ask for a time, the tool sets the video to that position and waits for the seek to finish before drawing the frame. The frame you get can differ from the exact timestamp by a few milliseconds, which is far below one frame at normal playback speeds. In practice you will not see the difference for a thumbnail or an illustration. The tool takes the time down to a hundredth of a second and the slider lets you scrub until the preview shows the exact picture you want, so if a single frame matters you can always nudge the position and check the preview before extracting. Each saved frame shows the time the tool actually used, so you can confirm it.
Yes, thirty frames per run in the two batch modes. Single frame mode is not affected because it only ever produces one image. The limit exists because every extracted frame is held in memory as a full image until you download it, and a large batch on a long clip can make the tab heavy or slow. Thirty frames is enough for a thumbnail set, a storyboard or a contact sheet, which is what most people need in one go. If you want more, extract thirty, download them, then run the tool again on another section. The counter under the mode selector tells you how many frames your current settings will produce, and it warns you when you have asked for more than thirty. Download all saves them one after another with a short pause between files, so your browser does not block the batch.
No. The tool saves the whole frame, scaled down if you choose a smaller width, but it cannot cut out a region from inside the picture. There is no crop box and no way to export only the top half or a face. The output size options of 1920, 1280, 854 and 640 pixels only change the overall width and the height follows to keep the aspect ratio, so nothing is trimmed off. If you need part of a frame, extract the full still first and crop it afterwards with the image cropper. That extra step keeps the extractor simple and predictable, and it means the frame you save is always a faithful copy of what the video holds rather than a resampled crop. A source narrower than the width you pick is kept at its own size instead of being enlarged.
Any format the browser can decode. MP4 with H.264 or H.265, WebM and most MOV files work in current browsers, and the tool reads the duration and resolution before it starts. If the browser cannot decode the container or codec, extraction reports that the frame could not be decoded instead of saving a broken image. That is the key difference from the video to GIF converter: this tool relies on the browser's own player, so it has no engine to download and starts instantly, but it also inherits the browser's codec support. When a clip will not decode, convert it to MP4 first with the video converter and then extract frames. AVI and some older containers may load on one browser and fail on another, so MP4 is the safest input. Extraction is free and there is no cap on how many times you use it.

Save your first frame now

Free, private and unlimited. No account needed.

Open Extract Frames