PlatformMCP server

MCP server

Agents

RenderKit ships a Model Context Protocol server so AI agents can screenshot, archive, and read the web without any wrapper code. Add one config block and your agent gets screenshot, pdf, and extract as native tools.

Install

Add RenderKit to your MCP client config (Claude Desktop, Claude Code, Cursor, …). The server runs via npx — nothing to install globally.

mcp config

{
  "mcpServers": {
    "renderkit": {
      "command": "npx",
      "args": ["-y", "renderkit-mcp"],
      "env": { "RENDERKIT_API_KEY": "rk_live_…" }
    }
  }
}

Exposed tools

Tool	Maps to	Use it for
`screenshot`	`POST /v1/screenshot`	Let the agent see a page exactly as a browser renders it.
`pdf`	`POST /v1/pdf`	Archive a page or generate a document.
`extract`	`POST /v1/extract`	Read clean, token-bounded markdown into the context window.

Why it fits agents

Extraction returns token-aware chunks, and every error code is machine-readable with retry hints — so an agent can recover on its own instead of surfacing a stack trace to a human.

Example prompt

to your agent

Screenshot https://stripe.com full-page, then extract the
pricing page to markdown and summarize the tiers.