AI Chat - Custom Branded Deployment

A minimal, clean deployment of Hanzo Chat with AI branding and custom tools.

Quick Start

# Initialize environment (copy template if missing)
cp .env.example .env

# Start AI Chat
make up

# View logs
make logs

# Stop
make down

AI Chat will be available at http://localhost:3081

Structure

chat/
├── .env                 # Branding and configuration
├── docker-compose.yml   # Main services
├── compose.prod.yml     # Production overrides
├── Makefile            # Simple commands
├── branding/           # Custom assets
│   ├── logo.svg
│   └── favicon.svg
├── tools/              # Custom MCP tools
│   └── financial/      # Financial analysis tools
├── models.yaml         # Custom model endpoints
└── data/               # Persistent storage

Configuration

Branding (.env)

Copy the environment template and update values:

cp .env.example .env

# Brand Configuration
BRAND_NAME=AI
BRAND_TITLE=AI Chat
BRAND_DESCRIPTION=AI-Powered Financial Document Analysis
BRAND_PRIMARY_COLOR=#6B46C1

# API Keys - Uncomment and add your keys
# OPENAI_API_KEY=sk-...
# ANTHROPIC_API_KEY=sk-ant-...

# Or use local API
# LOCAL_API_URL=http://localhost:4000/v1

Important: Don't leave API keys empty (OPENAI_API_KEY=) as this will override system environment variables. Either set a value or comment out the line.

Custom Models (models.yaml)

endpoints:
  - name: "AI Finance Models"
    baseURL: "${LOCAL_API_URL}"
    models:
      - ai-financial-v1
      - ai-credit-v1

Custom Tools (tools/)

Place MCP tool configurations in the tools directory:

tools/
├── financial/
│   ├── config.json
│   └── server.js
└── credit/
    ├── config.json
    └── server.js

Features

• Dynamic Branding: All branding controlled via BRAND_* environment variables
• Custom Tools: Mount your own MCP tools in /tools
• Custom Models: Configure additional model endpoints
• Demo Mode: Automatic demo login on localhost
• Production Ready: Includes production compose file with SSL, monitoring

Demo Login

When DEMO_MODE=true, SHOW_DEMO_LOGIN=true, and running on localhost:

• Email: ${DEMO_EMAIL}
• Password: ${DEMO_PASSWORD}

Production Deployment

# Configure production domain
echo "DOMAIN=chat.ai.ai" >> .env.prod

# Start with production config
make prod

Testing

# Run tests to verify branding and tools
make test

This runs:

• Branding verification
• Custom tool availability
• Model endpoint connectivity
• Demo login functionality

Customization

1. Branding: Update BRAND_* variables in .env
2. Tools: Add MCP servers to tools/
3. Models: Configure endpoints in models.yaml
4. Assets: Replace files in branding/

Requirements

• Docker & Docker Compose
• API keys (OpenAI/Anthropic) OR local LLM endpoint

Support

• Documentation: https://ai.ai/docs
• Issues: https://github.com/hanzoai/chat/issues