SnapRender Integrations

You can connect to SnapRender's MCP server to capture website screenshots from any MCP client without installing local software. This server-compliant interface lets you initialize a session, request screenshots in multiple formats, and reuse sessions across requests, making it easy to integrate SnapRender into web apps, automation workflows, or AI agents.

How to use

You interact with the SnapRender MCP server by connecting through a standard MCP client. Use the remote MCP endpoint to initialize a session, then send requests to capture screenshots or query usage. The server supports taking screenshots in PNG, JPEG, WebP, or PDF and lets you check cached results or review usage statistics. If you are building an integration, you will typically configure the MCP client to point at the remote endpoint and provide an API key in the request headers.

How to install

Prerequisites: you need Node.js or Python if you plan to run a local MCP server, or you can use a remote MCP server directly from your MCP client.

If you want to run a local MCP server, use the npm-based local setup demonstrated in the examples.

Additional sections

MCP endpoints and tools are designed to be used by any MCP client. The remote server supports transport via Streamable HTTP and provides the following tools and prompts for use in your integrations.

Configuration for clients (examples)

{
  "mcpServers": {
    "snaprender": {
      "type": "streamable-http",
      "url": "https://app.snap-render.com/mcp",
      "headers": {
        "Authorization": "Bearer sk_live_your_key_here"
      }
    }
  }
}

Local MCP server (npm) example

If you prefer running a local MCP server via stdio transport, you can use the npm-based setup shown here.

{
  "mcpServers": {
    "snaprender": {
      "command": "npx",
      "args": ["-y", "snaprender-mcp"],
      "env": {
        "SNAPRENDER_API_KEY": "sk_live_your_key_here"
      }
    }
  }
}

Remote vs Local usage

Remote (hosted) usage requires no local installation beyond your MCP client. Local usage runs the MCP server on your machine using stdio transport, typically via npx, and requires Node.js.

MCP Tools

The server exposes the following tools for your MCP client to call: take_screenshot, check_screenshot_cache, and get_usage. It also defines prompts such as screenshot_website and compare_devices to guide your screenshot workflows.

Direct API patterns you can follow

Use the remote MCP endpoint to initialize a session, then issue subsequent requests to take screenshots or check cache. You can reuse a session by including the session identifier in your subsequent requests.

Security and keys

Protect your API key and pass it in the request headers as Authorization: Bearer <your_key> or via X-API-Key depending on client capabilities.

Notes

The server supports multiple formats for output: PNG, JPEG, WebP, and PDF. You can control viewport size, device emulation, dark mode, and options to block ads or cookie banners as part of the screenshot parameters.

Available tools

take_screenshot

Capture a screenshot of a website and return it in PNG, JPEG, WebP, or PDF formats.

check_screenshot_cache

Check if a screenshot is cached without counting against quota.

get_usage

Retrieve screenshot usage statistics.

screenshot_website

Prompt used to initiate a website screenshot workflow.

compare_devices

Prompt used to compare screenshots across different device profiles.