u-he-preset-randomizer

Provides an MCP server to converse about, generate, and merge u-he presets via natural language and scripted workflows.
  • typescript

22

GitHub Stars

typescript

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "fannon-u-he-preset-randomizer": {
      "command": "u-he-mcp-server",
      "args": []
    }
  }
}

You can generate and tailor u-he synth presets by using a powerful command-line tool that analyzes your existing presets and creates new ones through controlled randomness and merging. This MCP-enabled server lets you interact with your preset collection via natural conversation, perform randomizations, and construct hybrid presets with flexible control over how much variation you introduce.

How to use

After you install and run the tool, you can use the MCP interface to ask for presets, explanations, and variations. You can generate fully random presets based on real value distributions from your library, randomize an existing preset with a specified randomness level, or merge several presets with random weighting. You can also narrow the pool of presets used for generation by folder, category, author, favorites, and other metadata. When using the MCP server, you can start conversations like: “Show me Diva presets” or “Create 5 variations of the ‘HS Greek Horn’ preset with 20% randomness.” The generated patches will be placed under a RANDOM subfolder inside your selected synth preset directory.

How to install

Prerequisites: ensure Node.js is installed on your system. You will run the tool from the terminal or command prompt.

Install and run in interactive mode with the latest version using npx. This will download and execute the tool interactively.

npx u-he-preset-randomizer@latest

MCP Server Interface

In addition to the command-line interface, you can use a Model Context Protocol (MCP) server to interact with your presets through AI assistants. The MCP server is exposed via a local stdio command that you run on your machine.

To enable the MCP server for conversational access, you can configure your AI assistant to start the server using the provided runtime command.

Example configuration you can add to your assistant setup (local MCP server):

{
  "mcpServers": {
    "uhe_preset": {
      "command": "u-he-mcp-server",
      "args": []
    }
  }
}

Tips & Tricks

Leverage the preset browser’s filters to narrow results by category, tags, or author. You can export a subset as a .uhe-fav file and use that as your narrow-down selection in the CLI. When merging presets, you can provide the same preset multiple times to adjust relative influence scores.

CLI Arguments / Configuration

If you are unsure which values to use, run the tool in interactive mode and rely on auto-complete suggestions.

Key options you will encounter include choosing the synth, number of presets to generate, the randomness level, and merging presets with explicit merge targets. You can also adjust how stable the results are and how binary sections are treated.

The tool supports both interactive usage and command-line arguments for automation. See examples in the usage section for concrete syntax.

Soundsets

Curated soundsets with generated presets are available as examples in the repository, demonstrating varied results across Diva and Repro synths. These soundsets illustrate how generated presets can be categorized, named, and tagged for easy reuse.

Developer Guide

To run locally in development mode, ensure Node.js and Bun are installed. You can install dependencies, build, and run tests or development servers as described.

Command sequence for development workflow includes installing dependencies, building, and starting the CLI in development mode. You can also run tests to verify behavior.

npm i
npm run build

# Run the CLI in development mode (using Bun):
bun run dev --synth Diva --amount 3

# Run tests:
bun test

# Maintenance scripts
npm run clean      # remove dist output and incremental build cache
npm run build:dist # emit declarations + source maps for npm publish (runs automatically via prepublishOnly)

Help / Feedback

If you need guidance or encounter issues, open a support thread or issue with clear steps to reproduce and, if possible, include the exact command you used and the synth you targeted.

Available tools

Search presets

Interactively search and browse your preset library through natural language queries to discover available synths, presets, and categories.

Explain presets

Get detailed explanations of what makes a preset unique, including parameter usage and sonic characteristics.

Generate random presets

Create fully random presets or variations based on real values and distributions from your library.

Merge presets

Combine multiple presets with random weighting to produce hybrid variations.

Stateful sessions

Maintain context across interactions, allowing you to select a synth once and explore related presets without reselecting.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational