Send a public URL or up to 200,000 characters of HTML. Get back PNG, JPEG (quality 1–100), WebP or PDF. Images can be the viewport, the full page, one element or a clipped region, at 1× to 3× pixel density.

Yes. Download mcp.mjs, then add it to your MCP client as a command: node mcp.mjs with SHOTS_API_KEY set to your key. The agent gets a take_screenshot tool. Each successful capture uses one credit.

Shots focuses on the common needs: clean captures, the main customization options, and simple prepaid credits with no subscription. It does not offer custom JavaScript, login cookies, proxies, caching, storage uploads or webhooks. If you need those, a larger hosted service may suit you better.

The page as rendered in a fresh Chromium browser, at your chosen viewport and capture delay. Full-page mode includes content below the fold. Bounded scrolling helps load images below the fold; very long or infinite pages may still be incomplete. You can add custom CSS and hide elements. You cannot submit login credentials, cookies or custom JavaScript.

No screenshot files are retained. The API returns the image directly; download it or save it in your own storage. We retain account details, credit balances, API key hashes and request metadata to operate the service.

No. Pages requiring login, CAPTCHA or unsupported network requests may not render. Internal networks, local addresses and non-standard ports are blocked. Capture only pages you are authorized to access.

A reserved credit is returned when rendering fails. A successful capture is charged once, even if you disconnect before downloading it. Duplicate successful API calls each use a credit. An error screenshot returned with a successful HTTP response from the destination may still count as a capture.

The private beta starts with trial credits. When paid packs are available, they are one-time purchases through Stripe, with no recurring subscription. Purchased credits are added after the payment is confirmed.