ESP32 Project Generator & Control Platform

An AI-powered platform that helps you create, manage, and control ESP32 IoT projects using natural language.

Video Demo Click to watch the demo video (7 min): 5 min project walkthrough + 2 min Nosana deployment

🤖 AI Agents Overview

This project leverages three specialized AI agents to handle different aspects of IoT project creation and management:

1. Project Planner Agent

• Understands natural language project descriptions
• Designs IoT project architecture including components, sensors, and outputs
• Plans MQTT topics and communication flow
• Generates project metadata and configuration

2. Code Generator Agent

• Takes project plans and generates ESP32 firmware code
• Handles WiFi, MQTT, and sensor integrations
• Implements automation rules and trigger responses
• Generates well-documented, ready-to-flash code

3. Triggering Agent

• Manages real-time device control through voice or text
• Handles natural language commands and converts them to MQTT messages
• Waits for device acknowledgments
• Provides voice feedback on command execution

🛠️ Tools & Features

Memory System (Redis)

• Persistent agent memory using Redis
• Stores project history and context
• Initialization via initial_memory.ts
• Memory updates through agent-memory-update.ts

MQTT Communication

• Real-time device control using MQTT
• Topic publication with acknowledgment waiting
• Automated reconnection and error handling
• Bi-directional communication with ESP32 devices

🚀 Local Setup

1. Clone & Install

git clone https://github.com/YOUR_USERNAME/agent-challenge
cd agent-challenge
bun install

2. Environment Setup

cp .env.example .env
# Edit .env with your configuration

3. Start Development Servers

# Terminal 1: Start Next.js UI
bun run dev:ui

# Terminal 2: Start Mastra Agent Playground
bun run dev:agent

The UI will be available at http://localhost:3000
Mastra Playground at http://localhost:4111

📦 Nosana Deployment

1. Build Docker Image

docker build -t yourusername/agent-challenge:latest .
docker push yourusername/agent-challenge:latest

2. Deploy on Nosana

• Open Nosana Dashboard
• Edit nosana_mastra_job_definition.json:

  {
    "image": "yourusername/agent-challenge:latest",
    "ports": ["3000:3000", "4111:4111"]
  }

• Submit job and monitor deployment

🎥 Video Overview

Video Demo The project demo video covers:

1. (0:00-5:00) Project Walkthrough

   • Agent architecture and interactions
   • UI demonstration
   • IoT project creation flow
   • Real device control demo

2. (5:00-7:00) Nosana Deployment

   • Docker image creation
   • Nosana dashboard walkthrough
   • Live deployment demonstration

🔗 Links

• Twitter Demo
• Nosana Dashboard
• Docker Image

📄 License

MIT License - See LICENSE for details

# Fork this repo on GitHub, then clone your fork
git clone https://github.com/YOUR-USERNAME/agent-challenge

cd agent-challenge

cp .env.example .env

pnpm i

pnpm run dev:ui      # Start UI server (port 3000)
pnpm run dev:agent   # Start Mastra agent server (port 4111)

Open http://localhost:3000 to see your agent in action in the frontend. Open http://localhost:4111 to open up the Mastra Agent Playground.

Step 2: Choose Your LLM for Development (Optional)

Pick one option below to power your agent during development:

Option A: Use Shared Nosana LLM Endpoint (Recommended - No Setup!)

We provide a free LLM endpoint hosted on Nosana for development. Edit your .env:

# Qwen3:8b - Nosana Endpoint
# Note baseURL for Ollama needs to be appended with `/api`
OLLAMA_API_URL=https://3yt39qx97wc9hqwwmylrphi4jsxrngjzxnjakkybnxbw.node.k8s.prd.nos.ci/api
MODEL_NAME_AT_ENDPOINT=qwen3:8b

If it goes down, reach out on Discord

Option B: Use Local LLM

Run Ollama locally (requires Ollama installed):

ollama pull qwen3:0.6b
ollama serve

Edit your .env:

OLLAMA_API_URL=http://127.0.0.1:11434/api
MODEL_NAME_AT_ENDPOINT=qwen3:0.6b

Option C: Use OpenAI

Add to your .env and uncomment the OpenAI line in src/mastra/agents/index.ts:

OPENAI_API_KEY=your-key-here