Skip to main content
Glama

A Model Context Protocol server for Scrapybara. This server enables MCP clients such as Claude Desktop, Cursor, and Windsurf to interact with virtual Ubuntu desktops and take actions such as browsing the web, running code, and more.

Prerequisites

Related MCP server: taw-computer

Installation

  1. Clone the repository:

git clone https://github.com/scrapybara/scrapybara-mcp.git
cd scrapybara-mcp
  1. Install dependencies:

pnpm install
  1. Build the project:

pnpm build
  1. Add the following to your MCP client config:

{
  "mcpServers": {
    "scrapybara-mcp": {
      "command": "node",
      "args": ["path/to/scrapybara-mcp/dist/index.js"],
      "env": {
        "SCRAPYBARA_API_KEY": "<YOUR_SCRAPYBARA_API_KEY>",
        "ACT_MODEL": "<YOUR_ACT_MODEL>", // "anthropic" or "openai"
        "AUTH_STATE_ID": "<YOUR_AUTH_STATE_ID>" // Optional, for authenticating the browser
      }
    }
  }
}
  1. Restart your MCP client and you're good to go!

Tools

  • start_instance - Start a Scrapybara Ubuntu instance. Use it as a desktop sandbox to access the web or run code. Always present the stream URL to the user afterwards so they can watch the instance in real time.

  • get_instances - Get all running Scrapybara instances.

  • stop_instance - Stop a running Scrapybara instance.

  • bash - Run a bash command in a Scrapybara instance.

  • act - Take action on a Scrapybara instance through an agent. The agent can control the instance with mouse/keyboard and bash commands.

Contributing

Scrapybara MCP is a community-driven project. Whether you're submitting an idea, fixing a typo, adding a new tool, or improving an existing one, your contributions are greatly appreciated!

Before contributing, read through the existing issues and pull requests to see if someone else is already working on something similar. That way you can avoid duplicating efforts.

If there are more tools or features you'd like to see, feel free to suggest them on the issues page.

Available Tools

5 tools
actB

Take action on a Scrapybara instance through an agent. The agent can control the instance with mouse/keyboard and bash commands.

Inputs: instance_id (required), prompt (required), schema.

bashC

Run a bash command in a Scrapybara instance.

Inputs: instance_id (required), command (required).

get_instancesB

Get all running Scrapybara instances.

Inputs: no named parameters.

start_instanceA

Start a Scrapybara Ubuntu instance. Use it as a desktop sandbox to access the web or run code. Always present the stream URL to the user afterwards so they can watch the instance in real time.

Inputs: no named parameters.

stop_instanceC

Stop a running Scrapybara instance.

Inputs: instance_id (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.5/5.0

    Scored across 5 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose: act controls the instance via agent, bash runs commands, get_instances lists instances, start_instance creates one, and stop_instance terminates one. There is no overlap or ambiguity between these functions.

    Naming Consistency4/5

    Four tools follow a consistent verb_noun pattern (get_instances, start_instance, stop_instance, run_command would fit but bash is used instead). The tool 'act' deviates as a single verb, and 'bash' is a noun, but overall the naming is mostly consistent and readable.

    Tool Count5/5

    With 5 tools, this is well-scoped for managing Scrapybara instances. It covers core lifecycle operations (start, stop, list) and usage actions (act, bash), with each tool earning its place without bloat or thinness.

    Completeness4/5

    The toolset provides good coverage for instance management and interaction, including lifecycle (start, stop, list) and control (act, bash). A minor gap is the lack of a tool for configuring or modifying instances, but agents can work around this using bash or act.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      A desktop automation MCP server that enables AI agents to interact with Linux environments through screenshots, window inspection, and input simulation. It provides tools for mouse control, keyboard input, and screen capture using xdotool and XDG Desktop Portals.
      MIT
    • A
      license
      B
      quality
      D
      maintenance
      An MCP server that provides AI agents with a full Ubuntu desktop environment inside Docker, enabling them to perform complex computer tasks like browsing, coding, testing, and GUI automation.
      36
      13
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      A Model Context Protocol server that gives any AI full control over a Windows PC, enabling application control, mouse/keyboard automation, screen capture, browser automation, and more.
      67 PyPI
      6
      MIT