RezNet AI - Local MVP

Multi-Agent AI Collaboration Platform

RezNet AI is a Slack-like chat platform where you orchestrate teams of specialized AI agents to tackle complex work. Start with our developer-focused agents (Backend, Frontend, QA, DevOps, Orchestrator), or build your own team for any domain - marketing, legal, research, and more.

Status: Local MVP - Single User Development Environment

---

Features

• Multi-Agent System: Coordinate specialized AI agents for different tasks
• 20+ LLM Providers: Anthropic, OpenAI, Google, Ollama, Groq, OpenRouter, Bedrock, and more via pi-ai
• Real-time Chat Interface: Slack-like UI with WebSocket communication
• Multi-Agent Workflows: DAG-based task orchestration with parallel execution
• MCP Integration: Model Context Protocol for filesystem and GitHub operations
• Agent Memory: Semantic memory with pgvector for context awareness across conversations
• Task Tracking: Built-in task management and delegation
• Full-Stack TypeScript: Unified TypeScript stack (Next.js frontend + Fastify backend)

Architecture

┌─────────────────────┐
│   Next.js Client    │  <-- Chat UI
│   localhost:3000    │
└──────────┬──────────┘
           │ WebSocket + REST
┌──────────▼──────────┐
│   Fastify Server    │  <-- Message Router
│   localhost:8000    │      Agent Manager (pi-ai + pi-agent-core)
└──────────┬──────────┘
           │
┌──────────▼──────────┐
│  Pi Agent Instances │
│  • @orchestrator    │  <-- Coordinates tasks
│  • @backend         │  <-- Backend expert
│  • @frontend        │  <-- React/UI expert
│  • @qa              │  <-- Testing specialist
│  • @devops          │  <-- Infrastructure expert
└──────────┬──────────┘
           │
┌──────────▼──────────────┐
│  MCP Servers (Local)    │
│  • Filesystem (3001)    │
│  • GitHub (3002)        │
└─────────────────────────┘

Quick Start

Prerequisites

• Docker Desktop (for PostgreSQL + Redis)
• Node.js 18+
• Anthropic API Key (or OpenAI/Google/Ollama/Groq)

Installation

1. Clone the repository

git clone https://github.com/alexg-g/reznet-ai.git
cd reznet-ai

2. Run the setup script

./scripts/setup.sh

This will:

• Check prerequisites
• Configure environment variables
• Start Docker services (PostgreSQL + Redis)
• Install backend dependencies (npm)
• Install MCP server dependencies
• Optionally set up the frontend

3. Add your API keys

Edit .env and add your Anthropic API key:

ANTHROPIC_API_KEY=your-key-here

Optional: Add GitHub token for GitHub MCP server:

MCP_GITHUB_TOKEN=your-github-token

4. Start all services

./scripts/start.sh

5. Access the application

• Backend API: http://localhost:8000
• API Docs: http://localhost:8000/docs
• Frontend: http://localhost:3000

Available Agents

@orchestrator

Role: Team Lead & Project Orchestrator

Coordinates development tasks, breaks down complex requirements, and delegates to specialist agents. Plans multi-agent workflows with DAG dependency resolution.

@backend

Role: Senior Backend Engineer Specializes in APIs, databases, and server-side logic.

@frontend

Role: Senior Frontend Developer Expert in React, Next.js, TypeScript, Tailwind CSS, and modern UI/UX.

@qa

Role: QA Engineer & Testing Specialist Writes comprehensive tests, finds edge cases, ensures code quality.

@devops

Role: DevOps Engineer Manages infrastructure, Docker, CI/CD, deployment, and monitoring.

Usage Examples

Via Chat Interface

Open http://localhost:3000 and interact with agents:

# Direct agent mention
@backend How do I implement JWT authentication?

# Orchestrated workflow
@orchestrator Build a user registration feature with email verification

# Context reset
/clear

Slash Commands:

• /clear - Reset channel context (agents won't see previous messages)

Via API

# Invoke an agent directly
curl -X POST http://localhost:8000/api/agents/invoke \
  -H "Content-Type: application/json" \
  -d '{
    "agent_name": "@backend",
    "message": "Explain how to implement JWT authentication",
    "context": {}
  }'

# Clear channel context
curl -X POST http://localhost:8000/api/channels/{channel_id}/clear \
  -H "Content-Type: application/json"

Via WebSocket

Connect to ws://localhost:8000/ws and send messages that mention agents like @backend, @frontend, etc.

Project Structure

reznet-ai/
├── backend-ts/            # TypeScript backend (Fastify)
│   ├── src/
│   │   ├── agents/       # Pi Agent instances + specialists
│   │   ├── db/           # Drizzle ORM schema + connection
│   │   ├── llm/          # pi-ai wrapper + error handling
│   │   ├── memory/       # Semantic memory (pgvector)
│   │   ├── routes/       # Fastify REST routes
│   │   ├── websocket/    # Socket.IO handlers
│   │   ├── workflows/    # DAG workflow engine
│   │   ├── config.ts     # Zod-validated config
│   │   └── index.ts      # Fastify entry point
│   ├── package.json
│   └── vitest.config.ts
│
├── frontend/              # Next.js frontend
│   ├── app/              # App Router pages
│   ├── components/       # React components
│   └── lib/              # Zustand store, utilities
│
├── mcp-servers/           # MCP servers
│   ├── filesystem/       # File system operations
│   └── github/           # GitHub integration
│
├── data/                  # Local data storage
├── scripts/              # Automation scripts
└── docker-compose.yml    # Docker services (PostgreSQL + Redis)

Development

Running Services Individually

Backend:

cd backend-ts
npm run dev

Frontend:

cd frontend
npm run dev

MCP Servers:

cd mcp-servers/filesystem && npm start
cd mcp-servers/github && npm start

Database Access

# PostgreSQL
docker exec -it reznet-postgres psql -U postgres -d reznetai_local

# Redis
docker exec -it reznet-redis redis-cli

Running Tests

cd backend-ts
npm test

Stop All Services

./scripts/stop.sh

Reset Database

./scripts/reset.sh

Troubleshooting

Port Already in Use

lsof -i :8000  # Find process
kill -9 <PID>  # Kill it

Database Connection Failed

docker-compose restart postgres
docker-compose logs postgres

Agent Not Responding

1. Check API key is set in .env
2. Check backend logs: tail -f logs/backend.log
3. Verify LLM provider is accessible

API Documentation

Interactive API docs available at:

• Swagger UI: http://localhost:8000/docs

Roadmap

Current (v1.0 - Local MVP)

• Multi-agent system with 5 specialist agents
• Real-time WebSocket communication
• Multi-LLM support (20+ providers via pi-ai)
• MCP integration (Filesystem, GitHub)
• Multi-agent workflow orchestration (DAG)
• Semantic memory with pgvector
• Task management

Next Steps

• Custom agent creation UI
• Workflow visualization
• Code execution sandbox
• Agent template library

Future

• Multi-user support
• Cloud deployment
• Agent marketplace
• Usage analytics

Contributing

We welcome contributions! See CONTRIBUTING.md for detailed guidelines.

How to Contribute

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

Community

• Issues: GitHub Issues - Bug reports & feature requests
• Discussions: GitHub Discussions - Questions & ideas
• Changelog: Check GitHub Releases

License

MIT License

Acknowledgments

• Uses Pi (pi-ai, pi-agent-core) for LLM abstraction and agent runtime
• Uses Model Context Protocol (MCP) for tool access
• Uses pgvector for semantic memory
• Powered by Anthropic Claude, OpenAI, Google Gemini, Ollama, Groq, and more

---

Built for developers, by developers