Chrome extension to save any image from the web as PNG, JPG, or WebP. Fast client-side conversion via Canvas API — no uploads, no servers, 100% private.
AVIF was removed in v1.2.0: Chrome's
canvas.toBlob()cannot encodeimage/avif, so the option could never work client-side.
Install from the Chrome Web Store
Edge Add-ons / Opera Add-ons — not published there yet. The extension is Manifest V3 and the same ZIP works in both browsers, so until then use the manual install below.
- Download the latest
save-image-as-type-<version>.zipfrom Releases, or clone this repo - Unzip it
- Open
chrome://extensions/ - Enable Developer mode (top right)
- Click Load unpacked
- Select the unzipped folder — or, if you cloned the repo, the
extension/folder
Manually installed extensions do not auto-update; repeat the steps to upgrade.
- Right-click context menu on any image: Save as PNG, JPG, WebP — plus a Save as default format item that skips the format submenu
- Client-side conversion via Canvas API (no server, no uploads)
- Quality sliders for lossy formats (JPG, WebP); PNG is always lossless
- Smart transparency handling (white background for JPG)
- Works on SVG,
data:URLs,blob:URLs, cross-origin and cookie-protected images - Keeps the original filename and swaps only the extension
- Always opens the "Save As" dialog, so you choose the destination
- Refuses to silently save the wrong format: if the encoder falls back, you get an error
- 100-megapixel guard read from the image header, so a decode bomb is rejected before any decoder allocates memory for it
- Welcome page on first install
- Manifest V3, minimal permissions
- Open source
The extension adds one submenu to the image right-click menu — there is no other UI to learn:
Clicking the toolbar icon opens a small settings popup: a Default format dropdown (used by the "Save as default format" menu item) and JPG quality / WebP quality sliders. PNG has no slider because it is lossless.
- Right-click any image on a webpage
- Select Save Image As from the context menu
- Choose your format: PNG, JPG, or WebP
- Pick where to save — done!
Tip: Click the extension icon to adjust default format and quality settings.
| Format | Type | Best For |
|---|---|---|
| PNG | Lossless | Screenshots, graphics, transparency |
| JPG | Lossy (adjustable) | Photos, smaller file size |
| WebP | Lossy (adjustable) | Modern web, 25-35% smaller than JPG |
Image-format converters are a crowded category, and most of them either upload your image to a server or have gone unmaintained. This one is deliberately narrow:
- Conversion never leaves the machine. The image is decoded and re-encoded with the browser's own Canvas API. There is no backend, no upload endpoint, no analytics, and no account. See Privacy.
- Manifest V3. Works under Chrome's current extension platform rather than the retired MV2 one.
- Only formats that actually work. AVIF was removed in v1.2.0 once it was confirmed
that
canvas.toBlob()cannot encodeimage/avif— shipping the option would mean a menu item that always fails. - Handles the awkward sources, not just plain
<img>tags: SVG (includingwidth="100%"and viewBox-only files),data:URLs,blob:URLs, cross-origin images, and cookie-protected images behind a login. - Open source, MIT. Every line that touches your images is in this repo, and releases are built by CI from a tag rather than uploaded by hand.
- Fully localized into 7 languages — extension name, store description, context menu, settings popup, welcome page, and every error message.
extension/
├── manifest.json # Manifest V3 config
├── background.js # Service worker: context menus, fetch, download orchestration
├── offscreen.html/.js # Offscreen document for Canvas API image conversion
├── i18n.js # Applies chrome.i18n messages to [data-i18n] elements
├── welcome.html # Onboarding page shown on first install
├── popup/ # Settings popup (format, quality sliders)
├── icons/ # Extension icons (16/32/48/128)
└── _locales/ # Localization (en, es, pt_BR, de, fr, ja, ru)
Why offscreen document? Service workers cannot use DOM Canvas with toBlob() for all formats. The offscreen document provides a real DOM context for image conversion.
Flow: Right-click image → background.js fetches image blob → sends to offscreen.js for Canvas conversion → downloads converted blob via chrome.downloads.
| Permission | Why |
|---|---|
contextMenus |
Right-click menu items |
downloads |
Save converted images |
storage |
Remember quality settings |
notifications |
Show conversion error messages |
offscreen |
Create offscreen document for Canvas API conversion |
scripting |
Read blob: image URLs from the page that created them |
<all_urls> (host) |
Fetch images from any domain (required for cross-origin image download) |
Note: <all_urls> host permission is broad but necessary — the extension has to fetch the
image you right-clicked, and that image can live on any domain. scripting is used on one
narrow path only: reading a blob: image URL from the page that created it. Every declared
permission is actually used; none is reserved "for later".
- Update version in
extension/manifest.json - Commit the change
- Create and push a git tag:
git tag v1.0.1
git push && git push --tagsGitHub Actions will automatically build a ZIP, create a GitHub Release, and upload and publish the new version to the Chrome Web Store. The CWS credentials are already configured in this repository — the pipeline has been shipping releases this way since v1.2.0.
The publish-chrome job in .github/workflows/release.yml runs when the repository variable CWS_EXTENSION_ID is set. Required configuration (Settings → Secrets and variables → Actions):
| Kind | Name | Value |
|---|---|---|
| Variable | CWS_EXTENSION_ID |
Extension ID from the CWS dashboard URL |
| Variable | CWS_PUBLISHER_ID |
Publisher ID from Developer Dashboard → Publisher → Settings (required by CWS API v2) |
| Secret | CWS_CLIENT_ID |
OAuth client ID (Google Cloud Console) |
| Secret | CWS_CLIENT_SECRET |
OAuth client secret |
| Secret | CWS_REFRESH_TOKEN |
OAuth refresh token with chromewebstore scope |
See Chrome Web Store API docs for obtaining the OAuth credentials. The first submission must still go through the CWS dashboard UI; auto-publish handles updates.
Download the ZIP from GitHub Releases and upload manually:
| Store | Dashboard | Cost |
|---|---|---|
| Chrome Web Store | CWS Developer Dashboard | $5 one-time |
| Edge Add-ons | Partner Center | Free |
| Opera Add-ons | Opera Developer | Free |
The same ZIP works for all three stores (Manifest V3 compatible).
- Version lives in
extension/manifest.json("version"field) - Git tags must match: tag
v1.0.1requires manifest version1.0.1 - CI verifies the match and fails if they differ
Everything the user sees is localized in extension/_locales/: the extension name, the store
description, the context menu, the settings popup, the welcome page, and all error
notifications.
Static pages opt in declaratively — extension/i18n.js walks [data-i18n] elements and
replaces their text at load. The English wording stays in the HTML as the fallback, so a
locale missing a key degrades to English rather than to a blank element.
| Language | Code |
|---|---|
| English | en |
| Spanish | es |
| Portuguese (Brazil) | pt_BR |
| German | de |
| French | fr |
| Japanese | ja |
| Russian | ru |
Your images are never uploaded. Conversion runs on the browser's own Canvas API inside an
offscreen document: the image is decoded, drawn to a canvas, re-encoded with toBlob(), and
handed to chrome.downloads. There is no backend, no upload endpoint, no analytics, no
tracking, no account, and no remote code — the extension loads no external scripts, styles,
or fonts.
Two details stated plainly, because "collects nothing" deserves the fine print:
- Fetching the image contacts its own host. To convert an image the extension downloads it first, exactly like the browser did when it displayed it. If that request fails or comes back as non-image content (a login page), it retries once with credentials, so cookies you already have for that host are sent to that host. Nothing goes anywhere else.
- Settings use
chrome.storage.sync. Your default format and the two quality numbers ride Chrome's own sync to your other signed-in Chrome devices. That is three values —defaultFormat,jpgQuality,webpQuality— and nothing about which images you saved, which sites you visited, or who you are.
See PRIVACY.md for the full policy.
MIT