Summer launch MouseMux V3 just launched. We'll be polishing it all summer with free updates, so early adopters get 25% off with code SUMMER26. See pricing
AI agents · MCP

Give your AI its own cursor.

MouseMux hosts a local MCP server. Connect your AI, arm it, and it gets its own cursor to click and type in any Windows app, no API required. One agent can drive many apps at once, each through its own locked virtual user.

In beta. The MCP server is local-only (127.0.0.1), off by default, and gated behind an Arm switch.

Input Mapper AI · MCP 127.0.0.1:41760 Arm
MouseSource · vector
KeyboardSource · binary
SmootherProcess
GateProcess
+ AILLMAI · agent
MergeProcess
Virtual UserOutput · app 1
+ AIVirtual User 2Output · app 2
Agent · app 1
Agent · app 2
Your AI reads and edits the graph over MCP, then, once you arm it, drives virtual users - one per app, working in parallel
Vector Scalar Binary signals flowing through the graph
How it works

A node graph your AI can read, rewire and run

The Input Mapper turns MouseMux input into a live signal graph. MCP hands that graph to your AI, and an Arm switch decides when it can act.

The Input Mapper

The Input Mapper is a visual node graph inside MouseMux. Every device that connects becomes a live signal source, mouse, keyboard, pen and touch, that you wire through processing nodes and out to real or virtual users.

Three signal types flow through it: vector positions, scalar values like pressure and wheel, and binary key and button events. Over 130 nodes cover math, filters, gates, generators, coordinate transforms, screen reading with OCR and image match, and automation macros.

That makes it a genuinely powerful toolkit: any input can be routed to become any output, and you can spin up unlimited virtual users - driving normal Windows apps, isolated multi-seat browsers, and soon any app fully sealed in its own container.

The MCP server

The app hosts a local Model Context Protocol server, so any MCP-compliant AI can connect: Claude Desktop and Claude Code work out of the box, and so does any other assistant, coding agent or agent framework that speaks MCP. The AI reads the whole graph, builds and rewires nodes, and creates virtual users on the machine - as many as the task needs, one per app, each locked to its own window so parallel work stays isolated and no human's pointer is ever touched.

Reading and building is always safe. Acting, moving, clicking and typing, is gated behind an Arm switch that is off by default and turns red while live, and the server listens on 127.0.0.1 only. Inside the graph, an LLM node can call Claude, OpenAI or your own API when a step needs judgement.

What it's good for

  • Legacy and no-API apps, driven through the real UI
  • One agent driving many apps in parallel, one locked virtual user per app
  • An agent working as its own user, next to your team
  • Screen-aware steps with OCR and image matching
  • Repetitive and QA runs, with an LLM in the loop when needed
On the roadmap

Soon: sealed app containers

One app, fully isolated

A real app container: the app runs sealed off from the OS, the agent drives only that app, and it can't touch anything else on the machine. Today's window locking, upgraded to full containment.

Many agents in parallel

One agent already drives many apps at once; next, run several agents side by side, each in its own isolated container on one PC, with no crosstalk between them and nothing leaking to the OS.

Hand your AI a cursor

Get the Input Mapper app, connect your AI, and arm it when you're ready.