ZumiLabs OmniCapt

Capture one page or your whole site. All local. All private

Chrome Extensions

Productivity

Developer Tools

Hey Product Hunt — I'm Amith, the builder behind ZumiLabs, and today I'm launching OmniCapt. I built it to scratch a real itch: I needed to capture lots of screenshots for a site being developed, at different resolutions, across many pages on my sites — to check responsive designs both on my local machine and on internal test servers. Doing that by hand (resize, screenshot, repeat, for every page and every breakpoint) was painful. But the tools that promised to automate it mostly run as hosted crawlers — they need your site reachable from the public internet, so capturing internal or pre-prod environments would've meant poking holes in firewalls or whitelisting IPs just to take a screenshot. That's a security and privacy trade-off I wasn't willing to make, so I needed something fully local. And browser capture extensions weren't much better: many are really front-ends for a cloud service that upload your screen, make you sign in, and add a delay to something I do dozens of times a day or only save the image to the downloads folder without any control over naming convention. So I built OmniCapt to do all the work in the browser. Nothing you capture leaves your machine. A lot of it evolved by hitting walls. "Capture every open tab" started by just activating each tab and screenshotting it — until I realised activating a tab is a focus change that closes the extension popup and kills the File System Access handle, so nothing saved; that's why it now uses Chrome's DevTools Protocol to shoot tabs in place instead. Saving started on chrome.downloads too, but a 150-file batch means 150 download prompts — unusable — so it moved to the File System Access API with a single folder grant. And the biggest shift was positioning: I built it as a one-page screenshot tool, then realised the part nobody else did well — bulk, multi-resolution, entirely local — was the actual point. So I leaned the whole thing into "capture one page, or your whole site." What makes it different: Every screenshot mode — visible, region, and full-page scroll-and-stitch (everything below the fold, one click). Capture at scale is the part most tools don't do. Paste a list of URLs (or import a sitemap) and it visits and screenshots every one, unattended — one file per page. Or capture every tab you already have open, by window or tab group, without ever switching focus (it uses Chrome's DevTools Protocol to shoot each tab in place). Queue up a URL set and walk away. Then multiply that by every resolution — define Desktop/Tablet/Mobile once and it captures each page at every size, auto-named with width×height. A 50-URL set across 3 breakpoints is 150 screenshots from one click. This is the responsive-QA feature I built it for. Everything saves straight to a local folder (File System Access API) with no download prompts, using a filename structure you define: slashes auto-create sub-folders, so captures sort themselves by domain, path, date, or even metadata scraped from the page (e.g. {{domain}}/{{path}}/{{year}}/{{dom:h1}}.png). Each one can carry a JSON sidecar with the page's URL, title, resolution, timestamp and scraped metadata (description, keywords, author, OG tags), so your captures are self-documenting and searchable. That searchability pays off in the built-in file browser. Browse every capture as thumbnails inside Chrome, filter across all that metadata to find any shot in seconds, then pick any two and compare them side-by-side or with a slider-wipe overlay to spot what changed between two runs — without leaving the browser. It also records your tab to WebM (tab audio, mic, on-screen timer, click ripples, pause/mute, and a built-in clip merger), and lets you annotate a page — arrows, boxes, text, full undo/redo — before you capture. It's local-first by design: no server, no account, no telemetry — open your network tab and nothing leaves. There's an optional CloudSend hook to push selected captures to Jira, Slack, GitHub, Confluence, Miro, S3, Google Drive or Contentful, but it stays hidden until you set it up, your token lives only in your browser, and the only time anything is uploaded is if you turn it on and point it at your own endpoint. (A hosted ZumiLabs backend is coming soon — early access on request.) It's free on the Chrome Web Store. It's part of a broader push at ZumiLabs to build fast, local-first browser utilities that just work. I'd genuinely love feedback on the workflow, and to hear what capture pain points you'd want solved next.