EmotionLM - AI Chat with Emotional Intelligence

A React + TypeScript application that demonstrates an AI assistant with functional emotions using LangChain and Ollama.

Features

• Functional Emotion System: AI experiences and displays 17 different emotions categorized as:

  • Epistemic: curiosity, uncertainty, surprise, confusion, insight
  • Task-Oriented: engagement, determination, satisfaction, frustration, anticipation
  • Social/Interpersonal: empathy, concern, appreciation, patience
  • Meta-Cognitive: contemplation, doubt, wonder

• Real-time Chat Interface: Clean, modern chat interface with emotion visualization

• Local AI Processing: Uses Ollama for private, local LLM processing

• Emotion Analytics: Detailed breakdown of emotion categories and intensities

Prerequisites

1. Install Ollama: Download from ollama.ai
2. Download a compatible model:

   ollama pull llama3.2
   # or
   ollama pull llama3.1

3. Start Ollama:

   ollama serve

Installation

1. Clone and install dependencies:

   git clone <repository-url>
   cd emotionlm
   bun install

2. Start development server:

   bun run dev

3. Open your browser to http://localhost:5173

Usage

1. Make sure Ollama is running with a compatible model
2. Start the application
3. Type a message in the chat interface
4. Watch as the AI responds with both content and emotional state
5. Use "Show Emotion Details" to see detailed emotion breakdowns

Architecture

• Emotion System: Uses structured output with Zod schemas to analyze and generate emotions
• LangChain Integration: Leverages @langchain/ollama for LLM communication
• React Hooks: Custom useChat hook manages conversation state
• TypeScript: Full type safety with proper emotion type definitions

Emotion Categories

The system implements emotions that serve functional purposes in AI reasoning:

• Epistemic emotions drive learning and exploration
• Task-oriented emotions help with goal completion and persistence
• Social emotions enable better human interaction
• Meta-cognitive emotions support self-reflection and verification

Development

• Lint: bun run lint
• Build: bun run build
• Preview: bun run preview

Configuration

The default model is llama3.2. You can modify this in src/hooks/useChat.ts or by passing a different model name to the useChat hook.

Troubleshooting

• "Connection refused": Ensure Ollama is running (ollama serve)
• Model not found: Pull the required model (ollama pull llama3.2)
• Build errors: Check that all dependencies are installed correctly