Snapshot Site is Now Available Directly Inside Claude

Snapshot Site now connects directly to Claude through MCP.

If you use Claude on the web, Claude Code, or the Anthropic API, you can call screenshot, analyze, and compare from inside your workflow as native tools.

claude.ai — one click, no API key

Go to claude.ai/customize/connectors, open Add connector, and enter:

https://mcp.snapshot-site.com/mcp

Claude handles the hosted sign-in flow automatically. Once connected, screenshot, analyze, and compare are available as tools in any conversation.

No API key setup. No configuration files.

Claude Code

If you use Claude Code, add the server to your mcp.json:

{
  "servers": {
    "snapshot-site": {
      "type": "http",
      "url": "https://mcp.snapshot-site.com/mcp",
      "headers": {
        "x-snapshotsiteapi-key": "ss_live_xxx"
      }
    }
  }
}

The tools appear directly in your development workflow. Ask Claude to capture a page, run a visual diff, or analyze content without switching to a separate integration.

Anthropic API

If you are building with the Anthropic API, attach the Snapshot Site MCP server to any messages.create call using mcp_servers:

import anthropic

client = anthropic.Anthropic()

response = client.beta.messages.create(
    model="claude-opus-4-6",
    max_tokens=1024,
    mcp_servers=[
        {
            "type": "url",
            "url": "https://mcp.snapshot-site.com/mcp",
            "name": "snapshot-site",
            "authorization_token": "YOUR_BEARER_TOKEN",
        }
    ],
    tools=[
        {
            "type": "mcp_toolset",
            "mcp_server_name": "snapshot-site",
        }
    ],
    messages=[
        {
            "role": "user",
            "content": "Capture a screenshot of https://example.com and describe what you see."
        }
    ],
    betas=["mcp-client-2025-11-20"],
)

authorization_token is sent as a Bearer token to the MCP server. Use the token issued for your Snapshot Site account when authenticating this connection.

This lets Claude call screenshot, analyze, and compare mid-conversation without you wiring the underlying API requests yourself.

What the tools do

Tool	What it does
screenshot	Capture a live page as PNG, PDF, or HTML with full rendering control
analyze	Extract content, metadata, and quality signals from a page
compare	Diff two page states or images and return mismatch metrics

When this is useful

This setup works well when the workflow starts from an AI assistant rather than from a custom integration:

• research sessions that need live page captures
• QA reviews that compare staging and production visually
• content audits driven by natural language queries
• agent pipelines that need rendered output without a browser dependency

For the full hosted MCP overview and setup guidance, see the Snapshot Site MCP documentation.