mcpField
O
OSCAR

OSCAR

Local
createwithoscar.siteMedia & Design192Updated 2w ago

About OSCAR

createwithoscar connects an AI assistant such as Claude to OSCAR, a free, open-source editor and server for control interfaces. It is aimed at people running OSC, MIDI or DMX setups who need to diagnose problems or draft new control pages. The assistant reads the local OSCAR state and never sends signals, publishes or edits projects.

Key features

  1. 01Reads projects, widget settings and MIDI ports from local OSCAR
  2. 02Shows recent OSC, MIDI and DMX traffic for diagnosis
  3. 03Drafts interfaces validated by OSCAR's own checks
  4. 04Drafts are marked Draft for the user to review and publish
  5. 05Requests stay on the local computer

Use cases

  1. 01Finding out why a slider does not control a light or application
  2. 02Drafting a lighting or audio control page from a text description
  3. 03Building a fader and blackout button page for a lighting desk

MCP support

This is an MCP server that exposes OSCAR's read and draft tools to MCP clients such as Claude Desktop and Cursor, connecting to the local OSCAR instance with its address and token.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"trafalmejo-oscar": {
"command": "npx",
"args": [
"-y",
"createwithoscar"
]
}
}
}
Run commandnpx -y createwithoscar

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmcreatewithoscar1.0.2

Tools

11

Saved tool list · updated 2w ago

This tool takes no parameters.

Details

From the project README

createwithoscar

Connects an AI assistant to OSCAR, the free, open-source editor and server for control interfaces: pages of buttons, sliders and other controls that phones and tablets open, and that send OSC, MIDI and DMX to show software, instruments and lights.

With it, an assistant such as Claude can:

  • Find out why something does not respond. It reads the OSCAR running on your computer: its projects, what is published, every widget's settings, the MIDI ports, and what OSCAR lately received and sent.
  • Draft an interface for you. It writes one, OSCAR's own validators check it, and it appears in OSCAR under File > Open project or template with a Draft badge, for you to open, check and publish yourself.

It never sends anything to your rig, publishes, or changes a project.

Connect

OSCAR 2.1 or later must be running on the same computer, with the MCP pill in its top bar green.

Claude Code

claude mcp add oscar -- npx -y createwithoscar

Claude Desktop: download OSCAR.mcpb from createwithoscar.site and double-click it. Nothing else to install.

Cursor, VS Code, Windsurf, Codex, Gemini CLI and other MCP apps: add a server with the command npx and the arguments -y createwithoscar. In most apps that is:

{
  "mcpServers": {
    "oscar": { "command": "npx", "args": ["-y", "createwithoscar"] }
  }
}

If it does not start on Windows, use the command cmd with the arguments /c npx -y createwithoscar.

Then ask, for example:

  • "Build me an OSCAR interface with four faders and a blackout button for my lighting desk."
  • "My OSCAR slider is not moving the light. Can you see why?"

How it works

This is a small program with no dependencies. The assistant starts it and speaks MCP to it; it passes each request to OSCAR on this computer and nowhere else. It finds OSCAR through a file OSCAR writes when it starts (~/.oscar/mcp.json), which holds OSCAR's local address and a token that changes at every start. Switching the MCP pill off in OSCAR closes the door.

If OSCAR is not running, the assistant still sees the tools, and is told to ask you to start OSCAR.

Needs Node 18 or later (the Claude Desktop bundle uses Claude's own).

Licence

BSD-3-Clause, as OSCAR is. Source: github.com/trafalmejo/OSCAR.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.