UBC LLM Chat

A WordPress plugin that creates a chat interface between the website and various large language models via their APIs.

Description

UBC LLM Chat allows WordPress site administrators to integrate large language models (LLMs) like OpenAI's ChatGPT and Ollama into their websites. The plugin provides a chat interface that can be embedded on any post or page using a shortcode or block.

Features

• Admin settings interface for configuring LLM services
• Support for OpenAI (ChatGPT) and Ollama
• Shortcode for embedding the chat interface on posts and pages
• WordPress block for embedding the chat interface using the block editor
• Multiple chat conversations per user
• Configurable rate limits and conversation limits
• Streaming responses for real-time interaction

Requirements

• WordPress 5.8 or higher
• PHP 8.0 or higher
• OpenAI API key (for ChatGPT integration)
• Ollama server (for Ollama integration)

Installation

1. Upload the ubc-llm-chat folder to the /wp-content/plugins/ directory
2. Activate the plugin through the 'Plugins' menu in WordPress
3. Go to 'Settings > UBC LLM Chat' to configure the plugin

Usage

Shortcode

You can embed the chat interface on any post or page using the following shortcode:

[ubc_llm_chat llmservice="chatgpt" llm="gpt-4o-mini" minimum_user_role="subscriber" maxmessages="20" maxconversations="10" systemprompt="Always speak like a pirate" temperature="0.7" debug_mode="true"]

Block

You can also embed the chat interface using the UBC LLM Chat block in the block editor.

Configuration

LLM Services

• OpenAI (ChatGPT): Enable OpenAI integration and provide your API key
• Ollama: Enable Ollama integration and provide the URL to your Ollama server

Global Settings

• Global Rate Limit: Set the number of seconds between API requests
• Global Maximum Conversations: Set the maximum number of conversations a user can have
• Global Maximum Messages: Set the maximum number of messages per conversation
• Connection Timeout: Set the connection timeout in seconds
• Minimum User Role: Set the minimum user role required to access the chat interface
• Debug Mode: Enable debug mode to output debug information to the log file

Development

Hooks

The plugin provides the following action hooks:

• ubc_llm_chat_settings_saved: Fired after settings are saved
• ubc_llm_chat_api_connection_tested: Fired after an API connection test
• ubc_llm_chat_models_fetched: Fired after models are fetched from an API
• ubc_llm_chat_usage_updated: Fired when usage statistics are updated

The plugin provides the following filter hooks:

• ubc_llm_chat_available_llm_services: Filter the available LLM services
• ubc_llm_chat_available_models: Filter the available models for a service
• ubc_llm_chat_model_parameters: Filter the parameters for a specific model
• ubc_llm_chat_rate_limit: Filter the rate limit for a user
• ubc_llm_chat_max_conversations: Filter the maximum number of conversations for a user
• ubc_llm_chat_max_messages: Filter the maximum number of messages per conversation
• ubc_llm_chat_user_can_access: Filter whether a user can access the chat interface
• ubc_llm_chat_streaming_enabled: Filter whether streaming is enabled for a specific instance
• ubc_llm_chat_api_request: Filter the API request before it's sent
• ubc_llm_chat_api_response: Filter the API response before it's returned

License

This plugin is licensed under the GPL v2 or later.

Credits

This plugin uses the LLPhant library for LLM integration.