Claude Code

Give Claude Code eyes

Add Grabbit's hosted MCP server in one command. Claude screenshots any URL, to verify the frontend it just built, capture a live site, or feed a real render into its own vision. No local Chromium, no API key to paste.

Get your free test keySee the install

One command, then Claude can see the web

Grabbit is a hosted MCP server at mcp.grabbit.live. Add it over Streamable HTTP and sign in through the browser, with no key to copy around.

Quick add (OAuth, no key)
# Add Grabbit's hosted MCP server (Streamable HTTP)
claude mcp add --transport http grabbit https://mcp.grabbit.live

# Authenticate in your browser (no API key to paste):
/mcp        # select "grabbit" → Authenticate
Alternative (Bearer key)
# Prefer a key? Pass it as a header instead of the OAuth flow.
# A free test key (sk_test_…) returns placeholder renders at no cost.
claude mcp add --transport http grabbit https://mcp.grabbit.live \
  --header "Authorization: Bearer sk_live_..."

Want the config printed for you? Run npx grabbit.live mcp and paste it in. Full API reference lives at /screenshot-api.

The loop coding agents have been missing

Claude Code writes frontend code with no way to see the result. Wire in Grabbit and it can capture the page it just shipped, look at the rendered pixels, and fix what is off: spacing, overflow, a broken image, a layout that only breaks at a real viewport. Then screenshot again to confirm.

see → fix → verify
You, in Claude Code:
  "Screenshot https://my-feature.preview.vercel.app and check the
   hero spacing against the mockup."

Claude calls the grab tool, looks at the returned image, edits the
CSS, then screenshots again to confirm the fix. A real see → fix → verify loop.

Point it at a deployed preview or staging URL (Grabbit captures public URLs, not localhost). Because the render runs on hosted infrastructure, the same loop works inside CI and remote runners where there is no local browser to drive.

Captures your agent's own browser can't get

Most agents can already drive a headless browser. It works on simple pages and quietly fails on the ones that matter: cookie and consent walls, bot detection, login or paywall gates, and heavy client-rendered pages. Grabbit runs the capture on hosted rendering infrastructure instead, so the shot comes back clean.

Local headless browser

  • Needs Chromium installed in the agent's runtime
  • Blank or blocked on bot-walled and consent-gated pages
  • You own the timeouts, retries, and scaling
  • Breaks in CI, sandboxes, and remote runners with no browser

Grabbit

  • One call, nothing to install or run
  • Dismisses cookie and consent banners automatically
  • Handles lazy-loaded, JavaScript-heavy, and full-page captures
  • Gets many sites a local headless browser is blocked on

Honest bound: no service captures literally every site. A handful of hard targets behind aggressive IP blocking still resist. But the cookie walls, consent gates, and JS-heavy pages that blank out a local headless browser are exactly what Grabbit is built to get.

Flat pricing, built for agents that loop

When an agent captures a page on every step, per-call metering turns into anxiety. Grabbit is one flat rate: $50 per year for 25,000 captures ($0.002 each), with prepaid packs that stack on top and never reset. No monthly quota to forfeit, no usage cliff. Test-environment keys render free placeholders, so you can wire the agent up before adding a card.

See full pricingCompare providers

Claude Code FAQ

Do I need an API key to use Grabbit in Claude Code?
No. Grabbit's MCP server supports OAuth, so Claude Code signs you in through the browser with no key to paste. If you prefer a key, pass a Bearer token as a header instead. Either way, a free test key returns placeholder renders so you can wire it up before paying.
Does this need a local headless browser or Chromium?
No. Grabbit runs the capture on hosted rendering infrastructure. Nothing is installed in your environment, so it also works in CI, sandboxes, and remote runners where a local browser like Playwright or Puppeteer would have nothing to drive.
Can Claude Code screenshot localhost?
Not directly. Grabbit captures public URLs and blocks private and internal addresses (including localhost) with built-in SSRF protection. To verify local work, point it at a deployed preview or staging URL, or expose your dev server through a public tunnel.
How is this different from the Playwright MCP server?
Playwright MCP drives a real browser you run locally. It needs Chromium installed, you own the scaling and timeouts, and it can blank out on bot-walled or consent-gated pages. Grabbit is a hosted call: no browser to install, cookie and consent banners handled, and it gets many sites a local headless browser is blocked on. Use Playwright MCP when you need to click and type through a flow; reach for Grabbit when you just need a clean image of a URL.
What tools does the Grabbit MCP server add?
Five: grab (screenshot a URL), get_grab and list_grabs (fetch or browse captures), get_usage (credits and plan), and get_pricing (an honest, dated comparison against ScreenshotOne, Urlbox, Browserless, and others).
What does it cost?
A flat $0.002 per live capture on the $50/year plan (25,000 captures), with prepaid packs that never reset or expire. Test-environment captures are free. No per-call meter to watch while an agent loops.

Give Claude Code eyes in two minutes

Free test key, one install, and your agent can screenshot the web.

Get your free test keyAll agent integrations