dosbox-mcp
Bridges AI agents to a DOSBox emulator, enabling control of DOS programs via MCP tools for typing, screen reading, video capture, Lua scripting, and memory access.
README
dosbox-mcp
MCP server for dosbox-automation: drive a DOS machine from an AI agent.
dosbox-automation is a DOSBox fork with a built-in automation API. This bridge connects any MCP-capable agent runtime (Claude Code and friends) to a running emulator. The agent can type into DOS programs, read the screen, capture frames and video, run sandboxed Lua scripts, and inspect memory: everything the automation API offers, as MCP tools.
Quick start
-
Install dosbox-automation and start it with the API enabled. The manual covers setup for all platforms.
-
Add the bridge to your MCP client. For Claude Code:
claude mcp add dosbox -- uvx dosbox-mcpOr in a JSON MCP configuration:
{ "mcpServers": { "dosbox": { "command": "uvx", "args": ["dosbox-mcp"] } } }pipx install dosbox-mcporpip install dosbox-mcpwork just as well; the command isdosbox-mcpeither way.
That's it. The bridge finds the running emulator, checks what it can
do, and registers the matching tools. Ask your agent for bridge_status
to confirm the connection.
Configuration
Optional. Write a commented config file to tune the port, spawn an emulator on demand, or limit what an agent may do:
dosbox-mcp setup --init
The comments in the generated file explain each setting. Environment
variables DOSBOX_API_URL and DOSBOX_API_TOKEN override it. The
bridge talks to loopback addresses only; it will not reach out to a
remote machine.
What the agent gets
Tools for session control, screen text and captures, video recording, Lua scripting, keyboard and mouse input, and memory access. You choose how much reach an agent has, from read-only observation to full control.
An open protocol
The contract between this bridge and the emulator is written down in PROTOCOL.md and versioned independently. dosbox-automation is the reference implementation; any DOSBox variant that speaks the protocol works with this bridge.
In the box
Like the games it drives, this one ships with its paper. The feelies/
folder holds the manual, a reference card for every tool, an install
sheet, and a tech-support page.
Requirements
- Python 3.11 or newer
- A running dosbox-automation (0.84-da2 or newer) on the same machine
License
GPL-2.0-or-later. See LICENSE. The protocol specification (PROTOCOL.md) is licensed CC-BY-SA-4.0.
This project is developed with tooled assistance, but tested, reviewed and signed off by a human developer.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.