MCP server

Let Claude and other agents capture a page, compare two of them, and audit one for accessibility. The capture tools return the image itself, not a link to it — so the model can look at the page it just asked for.

Why that matters

Most screenshot integrations hand an agent a URL and stop there. The model knows a capture happened and cannot see it.

These tools return an MCP image block, so “screenshot this page and tell me what’s broken” finishes in one turn with no fetch step in between.

Install

No install step — your client runs it with npx. You need a getSnap API key; the free tier is 100 captures a month, no card.

Claude Code

claude mcp add getsnap --env GETSNAP_API_KEY=sk_live_... -- npx -y getsnap-mcp

Claude Desktop

In claude_desktop_config.json:

{
  "mcpServers": {
    "getsnap": {
      "command": "npx",
      "args": ["-y", "getsnap-mcp"],
      "env": { "GETSNAP_API_KEY": "sk_live_..." }
    }
  }
}

The same block works in any client that speaks MCP over stdio.

Tools

ToolCreditsReturns
capture_screenshot1The image, inline
capture_pdf1A link — a PDF cannot be shown to a model
compare_pages2Similarity score and the diff overlay, inline
audit_accessibility1WCAG violations by severity, with the elements
check_usage0Credits used and remaining

Ask for things in plain language

Three things worth knowing

Configuration

VariableRequiredDefault
GETSNAP_API_KEYyes—
GETSNAP_BASE_URLnohttps://api.getsnap.dev

Without a key the server exits immediately with a message saying where to get one, rather than starting and failing on the first call.

Free key, 100 captures a month

No card. Works with the MCP server, the REST API, and the Node and Python SDKs.

Get a free API key