Skip to main content
Glama

input-mcp

MCP server for collecting contextual user input. The server exposes a single collect_input tool that can collect:

  • Text input - simple text or JSON from the user

  • Image input - freeform drawing on a canvas

  • Pixel art - grid-based pixel art editor with color palette

When called, it launches an Electron window where the user can provide their input, then returns the result (as a file path for images) back to the MCP client.

Pixel Art Example

Quick Start

npm install
npm run build
npx @modelcontextprotocol/inspector node dist/src/server.js

Note: This package is also published on npm as @swairshah/input-mcp and can be used directly with npx @swairshah/input-mcp.

This opens a web interface where you can test the tool. Try calling collect_input with different parameters:

  • {"kind": "text"} - opens a text input window

  • {"kind": "pixelart", "gridWidth": 16, "gridHeight": 16} - opens a pixel art editor

  • {"kind": "image", "initialImage": "/path/to/image.png"} - opens an image editor with a starting image

Related MCP server: prompt-gui-mcp

Build and Test the UI Components

bun install first. obviously.

  1. Build the UI bundle (creates ui/dist/ assets used by Electron):

    npm run build:ui
    # or
    bun run build:ui
  2. Launch the Electron prompt helper directly (useful for smoke tests):

    bun run create

    This spawns the image/text prompt window with the default text spec.

  3. Test script:

    bunx tsx scripts/test-input.ts image

Testing with MCP Inspector

npm run build
npx @modelcontextprotocol/inspector node dist/src/server.js

The inspector provides a web UI to test the tool. Images are saved to ~/.cache/input-mcp/images/ and the tool returns the file path.

Using with Claude Desktop

Add to your Claude config (~/Library/Application Support/Claude/claude_desktop_config.json):

Using npm package (recommended):

{
  "mcpServers": {
    "input-mcp": {
      "command": "npx",
      "args": ["-y", "@swairshah/input-mcp"]
    }
  }
}

Using local clone:

{
  "mcpServers": {
    "input-mcp": {
      "command": "node",
      "args": ["/absolute/path/to/input-mcp/dist/src/server.js"]
    }
  }
}

Then ask Claude things like:

  • "Let me draw something for you" (opens drawing canvas)

  • "I want to create pixel art" (opens pixel art editor)

Dev

Project Structure

shared/       → Zod schemas, shared types, and error helpers
ui/           → Electron renderer (HTML/CSS/JS) and prompt modules
create.ts     → Launches the Electron window and normalises specs
server.ts     → MCP server definition for the `collect_input` tool
scripts/      → Ad-hoc utilities (`test-input.ts` for manual runs)

Development Workflow

  • Modify the renderer in ui/renderer.ts and module files under ui/modules/.

  • Add new input kinds by extending shared/types.ts and branching inside mount*Module helpers.

  • When iterating on the UI, run bun run create (or npx tsx scripts/test-input.ts image) to open a live window with the current spec.

License

MIT

Available Tools

1 tool
collect_inputCollect InputC

get image, text, or pixel art input from user. This is used to get contextual input from the user of different kinds.

Inputs: kind, initialImage, gridWidth, gridHeight, +3 more.

Tool Schema Changelog

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

Loading tool schema changelog…

    TDQS

    C2.8/5.0

    Scored across 1 tool

    Disambiguation5/5

    With only one tool, there is no possibility of ambiguity or overlap between tools. The tool's purpose is clearly defined and distinct by default.

    Naming Consistency5/5

    A single tool inherently has consistent naming, as there are no other tools to compare it against. The name 'collect_input' follows a clear verb_noun pattern.

    Tool Count2/5

    One tool is too few for most server purposes, as it severely limits functionality and scope. It feels thin and incomplete for handling user input in a robust way.

    Completeness2/5

    The server's purpose appears to be collecting user input, but with only one tool, there are significant gaps. For example, no tools for validating, processing, or managing different input types, making the surface severely incomplete.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      B
      quality
      D
      maintenance
      Collects user feedback with text and image support through an Electron app, allowing AI tools to gather and process user input with customizable prompts and multiple response options.
      1
      9 npm
      2
      Apache 2.0
    • A
      license
      A
      quality
      D
      maintenance
      Enables AI agents to prompt users for input directly within the chat UI using MCP's elicitation capabilities, allowing for interactive multi-turn conversations without leaving the conversation interface.
      1
      14 npm
      1
      MIT
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables AI agents to select Windows desktop UI elements, windows, or screen regions via mouse hover, then obtain context through UI Automation, screenshots, and local OCR for MCP-compatible clients.
      2
      MIT