Video to WebP Converter

Video to WebP: What This Tutorial Covers

This page pulls still frames out of a video and saves them as WebP images. It does not build an animated WebP — one run gives you either a single frame captured at a timestamp you choose, or a set of separate frames sampled across the clip. That distinction trips people up constantly, so it is worth saying twice: the output is a photo of a moment, not a moving image. If you want something that loops, use video to GIF or video to APNG instead. What follows covers picking the right frame, choosing between lossy and lossless, and the handful of things about video frames that make a still come out looking wrong.

How to Convert Video to WebP

1. Upload Your Video File: Drag a clip onto the page or use the file picker. MP4, MOV, MKV, WebM, AVI, WMV, FLV, MPEG, M4V, TS, MTS, VOB, 3GP and around two dozen more containers are accepted, and several files can be queued at once.
2. Choose Specific Frame or Multiple Screenshots: Under "Frame Selection", "Specific Frame" is preselected with "Time (seconds)" set to 0 — change that number to grab any moment you like. "Multiple Screenshots" instead samples the whole clip at a "Capture Rate" that defaults to "1 second per frame" and ranges from "0.1s seconds (single frame at 10fps)" up to "10 seconds per frame".
3. Set Quality Preset, Lossless and Image resolution: The Quality Preset dropdown opens on "Very High (Recommended)" and runs from "Lowest" to "Highest"; "Lossless?" defaults to "No (Recommended)"; Image resolution starts on "Keep original". There is also a "Specific file size" mode if you need the output under a fixed byte budget.
4. Convert and Download: Click Convert and save the image, or download a multi-frame set together. Files are uploaded over an encrypted connection, processed on our servers, and deleted automatically after a few hours. No sign-up, no watermark.

Walk-through: Specific Frame vs Multiple Screenshots

Both paths decode to a lossless intermediate first and only then encode WebP, so nothing is quantised twice.

Specific Frame seeks to the timestamp in "Time (seconds)" and decodes exactly one frame. The default of 0 means the very first frame of the file — which on a lot of real footage is a black frame, a fade-in, or a slate, and is the single most common reason people think the tool "returned a blank image". Type a real timestamp. Decimals work, so 12.5 is valid, and if the clip opens on a fade you usually only need to move a second or two in.

Multiple Screenshots samples the entire clip at the Capture Rate and writes each sample as its own WebP file. The dropdown is phrased in seconds per frame, so bigger numbers mean fewer images:

• "0.1s seconds (single frame at 10fps)" — ten images per second of source. Useful for finding the exact instant something happens; expect a lot of files from anything longer than a few seconds.
• "1 second per frame" (the default) — one image per second. A sensible contact sheet for a clip of a minute or two.
• "10 seconds per frame" — one image every ten seconds. The right choice for skimming a long recording or building chapter thumbnails.

Every screenshot in the set uses the same Quality Preset, Lossless setting and Image resolution. There is no per-frame control, and the frames are independent images — nothing stitches them back together.

Lossy or Lossless? Choosing the Right WebP Mode

The alpha behaviour is worth calling out because no other common format does it: WebP can pair a lossy RGB encode with a losslessly compressed alpha channel, which Google's own documentation describes as a combination "not available today with any of the existing image formats". So a frame with soft transparent edges keeps clean edges even in lossy mode.

WebP vs JPEG vs PNG for a Video Frame

If a tool downstream refuses WebP, video to JPG covers maximum compatibility and video to PNG covers lossless with transparency.

Common Errors and How to Fix Them

• "The image is black" — "Time (seconds)" is still on 0 and the clip opens on black or a fade. Enter a timestamp a few seconds in.
• "I expected an animation and got one picture" — This conversion produces still frames only. Use video to GIF or video to APNG for a moving result.
• "Text and thin lines look mushy" — Lossy mode at a middling preset. Set "Lossless?" to Yes for screen recordings and UI captures, or raise the Quality Preset to "Highest".
• "The frame looks horizontally squashed" — Your source has non-square pixels; DVD and broadcast standard-definition video stores a 720-pixel-wide frame that players stretch at display time. A still image has no field to carry that ratio, so set Width x Height explicitly — 854 × 480 for 16:9 NTSC, 1024 × 576 for 16:9 PAL.
• "There are comb lines across moving edges" — The source is interlaced, and no deinterlacing is applied anywhere in this pipeline. Each captured frame contains two fields taken moments apart. Deinterlace the video before uploading, or pick a frame where nothing is moving.
• "I got dozens of files when I wanted one" — "Multiple Screenshots" was selected. Switch back to "Specific Frame", or raise the Capture Rate to a larger seconds-per-frame value.

When This Doesn't Work

If what you actually need is an animated WebP, this is not the page — the pipeline routes still-image targets through a frame extractor and only GIF and APNG through the animated assembler. If the video is DRM-protected, no frame can be decoded at all. And if you are chasing one precise instant in a long recording, seeking to a timestamp inside a large file is slower and less exact than trimming first: cut the second you care about with the video cutter, then extract from the short result.

Frequently Asked Questions

Does this tool produce an animated WebP?

No. Still frames only. "Specific Frame" writes one WebP at the timestamp you set, and "Multiple Screenshots" writes a series of separate WebP images — each an independent file, never stitched into a moving image. The frame-rate-style wording in the Capture Rate dropdown describes how often the clip is sampled, not a playback rate. For a genuine animation, use video to GIF or video to APNG.

How much smaller is a WebP frame than the same frame as JPEG or PNG?

For a lossy still, Google's WebP compression study measured WebP at 25–34% smaller than JPEG at the same Structural Similarity index, with file-size ratios between 0.66 and 0.75 across their test sets and quality levels. For a lossless still, MDN puts lossless WebP at roughly 26% smaller than the equivalent PNG. Real savings vary with content: flat or simple frames compress much better than detailed photographic ones.

Will the WebP frame keep transparency from my video?

Only if the source carries an alpha channel. Frames are extracted with alpha intact, and a ProRes 4444 MOV whose pixels are 60% opaque produces a WebP with those same opacity values, corner-transparent areas included. Standard MP4 with H.264, AVI, MPEG-2 in a VOB and ordinary MOV recordings have no alpha, so their frames are fully opaque. WebP supports transparency in both its lossy and lossless modes.

Which frame do I get if I leave the time at 0?

The first decodable frame in the file. That is often not a useful image — many clips begin with black, a fade-up, a slate or a logo card — so treat 0 as a placeholder rather than a default worth keeping. Any decimal value is accepted, so you can land on a precise moment.

Can I set an exact output size in kilobytes?

Yes. Instead of Quality Preset, choose "Specific file size" under Image Compression and enter a number with Bytes, Kilobytes or Megabytes. There is an "Auto Scale" toggle alongside it that allows the dimensions to be reduced when the byte budget cannot be met by compression alone — useful when a page has a hard asset-size limit and the frame is high resolution.

Does the Quality Preset do anything when Lossless is set to Yes?

No. Lossless encoding reproduces the decoded frame exactly, so there is no quality level to trade away — the preset only governs the lossy encoder. Bear in mind that lossless is genuinely large for photographic content; Google's own figures put lossless alpha alone at about 22% more bytes than a quality-90 lossy encode, and lossless colour costs far more than that. Use it for screen captures and graphics, not for camera footage.

Does the extracted frame keep the video's rotation?

Yes. Rotation and flip are applied after extraction, with orientation metadata normalised first, so a clip recorded in portrait on a phone produces an upright still rather than a sideways one. If you also ask for a rotation or flip on the page, it is applied to every frame in the set.

What does Image resolution do to an extracted frame?

It scales the still before it is encoded. "Keep original" gives you the video's native frame size; Preset Resolutions, Width, Height, Width x Height and Resolution Percentage all shrink it. Presets set the height and let the width follow the source aspect, so use Width x Height when you need exact dimensions — including when you are correcting a non-square-pixel source.