Reze AI Agent

An intelligent AI assistant for Resend.com that sends emails via natural language, tracks delivery status, and provides API assistance using RAG.

Features

• Chat-only interface through natural conversation
• Email operations: send emails, check status, retrieve attachments
• RAG documentation powered by Memvid with Resend.com docs
• Real-time streaming responses
• GLM 4.7 AI model via z.ai (OpenAI-compatible)
• Multi-turn conversations with full context memory
• Username-based conversation isolation
• Markdown rendering for formatted responses
• Persistent SQLite storage for conversation history

Quick Start

# Install dependencies
pip install uv
uv sync

# Configure environment (edit .env with your API keys)
cp env.example .env

# Start server
uv run python main.py

# Visit chat interface at http://localhost:8000

Environment Variables

Required variables in .env:

# GLM 4.7 AI
GLM_API_KEY=your_glm_api_key
GLM_BASE_URL=https://open.bigmodel.cn/api/paas/v4
GLM_MODEL=glm-4.7

# Resend.com
RESEND_API_KEY=re_xxxxxxxxxxxx
RESEND_FROM_EMAIL=noreply@yourdomain.com

# Memvid RAG
MEMVID_FILE_PATH=./memory.mv2

# Database
DATABASE_URL=sqlite+aiosqlite:///./database.db

Usage Examples

Send Email via Chat

curl -N -X POST http://localhost:8000/api/chat/message \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Send an email to john@example.com welcoming him to our platform",
    "username": "user1"
  }'

Agent asks clarifying questions (subject, content), then sends email.

Check Email Status

curl -N -X POST http://localhost:8000/api/chat/message \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Check the status of that email",
    "conversation_id": "previous_conversation_id",
    "username": "user1"
  }'

Agent remembers context and checks delivery status.

Ask About Resend Features

curl -N -X POST http://localhost:8000/api/chat/message \
  -H "Content-Type: application/json" \
  -d '{
    "message": "What are the email size limits?",
    "username": "user1"
  }'

Agent uses RAG to answer from Resend documentation.

API Endpoints

Interactive API documentation available at: http://localhost:8000/docs

Method	Endpoint	Description
POST	/api/chat/message	Chat with AI agent (streaming)
GET	/api/chat/conversations/{username}	List user conversations
GET	/api/chat/conversations/{id}/history	Get conversation history
DELETE	/api/chat/conversations/{id}	Delete conversation
GET	/health	Health check endpoint

Tech Stack

• Backend: FastAPI, Python 3.12, SQLAlchemy (async)
• AI: PydanticAI, GLM 4.7 (z.ai), OpenAI-compatible API
• Knowledge Base: Memvid (semantic search)
• Database: SQLite + aiosqlite
• Frontend: Vanilla JavaScript with marked.js

Populate Documentation

# Populate Memvid with Resend.com documentation
uv run python scripts/populate_memvid.py

Architecture

User Request
    ↓
FastAPI Endpoint
    ↓
AI Agent (GLM 4.7) + Memvid RAG
    ↓
Resend.com API
    ↓
Streaming Response

Project Structure

ai-resend-agent/
├── src/
│   ├── api/          # FastAPI routes and schemas
│   ├── services/      # AI, RAG, Resend, conversation
│   ├── tools/         # PydanticAI tools for Resend operations
│   ├── database/      # SQLAlchemy models and session
│   └── config.py     # Configuration
├── scripts/
│   └── populate_memvid.py
├── main.py           # Application entry point
└── README.md

License

MIT License - see LICENSE for details.

Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Support

• Issues: https://github.com/Nneji123/reze-agent/issues
• Discussions: https://github.com/Nneji123/reze-agent/discussions