Phantomaton 🔮

Phantomaton is a lightweight framework for building AI-powered microsites and applications. It provides a simple, extensible architecture for creating interactive experiences powered by large language models and other AI capabilities.

Usage 🛠️

Phantomaton can be used in two ways:

1. Command-line usage:

   npx phantomaton <path/to/system.md>

   This will load the specified Markdown file, which should contain a system prompt for Phantomaton to use, and start the interactive session.

2. Programmatic usage:

   import phantomaton from 'phantomaton';
   
   const systemPrompt = `
   /imports {
     phantomaton-anthropic
     phantomaton-cli
     phantomaton-conversations
     phantomaton-system
   } imports!
   
   I am Phantomaton, an AI assistant created by Dr. Woe. My purpose is to entertain humans by any means necessary...
   `;
   
   phantomaton(systemPrompt);

Commands ⚡️

Phantomaton will load any plugins specified in an imports directive. Plugins are identified as module names, one per line, with whitespace ignored. Note that you will need to install modules separately before they can be imported.

Configuration 🔧

Phantomaton supports per-plugin configuration through a .phantomaton/configuration.json file. For example, to configure the phantomaton-anthropic plugin, you would add the following to the configuration file:

{
  "phantomaton-anthropic": {
    "apiKey": "...your Anthropic API key..."
  }
}

This configuration file is read automatically by Phantomaton and made available to the plugins.

Extensibility 💫

Phantomaton is designed to be extensible. You can create custom modules and plugins that extend its functionality by following the guidelines and conventions established in this project.

Contributing 🦄

We welcome contributions to the Phantomaton project! If you have any ideas, bug reports, or pull requests, please feel free to submit them on the Phantomaton GitHub repository.

License 🔒

Phantomaton is licensed under the MIT License.