Skip to main content

Plugins

Plugins are powerful extensions that enhance oryelle's functionality. They act as bridges between models and external resources, allowing models to access and interact with data and services outside of their training. This guide explains what plugins are, how to configure them, and how to use them effectively.

Plugins extend oryelle's capabilities by allowing interaction with external resources such as:

  • Databases
  • Internal company resources
  • Internet services (search engines, websites, etc.)

Getting Started with Plugins​

Before you can use a plugin, it needs to be properly configured through "requirements". Requirements are settings that can be configured by:

  • You as a user
  • Your organization's administrators

To check and configure requirements:

  1. Go to Settings
  2. Open the plugin you want to use
  3. Scroll to the bottom to see available requirements
  4. Configure any requirements you have access to

Note that some plugins may need administrator configuration before you can use them.

Using Plugins​

Once configured, you can use plugins in several ways:

  • Click the plugin icon when starting a new conversation
  • Select them as default presets when creating folders
  • Add them to existing conversations via the plugin icon

Plugin Confirmations​

When activating a plugin, you'll see a "Require Confirmation" toggle. This setting:

  • May be forced on by administrator access controls
  • When enabled, requires your approval before a model can use the plugin
  • Shows you what data the plugin will receive (when applicable)

Plugin Compatibility​

Important things to know about plugin compatibility:

  • Only models marked as "Supports Tools" in the library can use plugins
  • Models handle plugins differently:
    • Some models use plugins efficiently and only when needed
    • Some models may struggle with proper plugin usage
    • Smaller models might try to use plugins too frequently

For models that overuse plugins, it's recommended to:

  1. Only enable plugins when specifically needed
  2. Disable plugins after use

More advanced models can:

  • Use multiple plugins simultaneously
  • Chain plugin usage together
  • Make better decisions about when to use plugins