ExperimentAI is a local-first AI-powered virtual science laboratory that enables students to perform physics, electrical, and astronomy experiments with real-time AI assistance, offline capability, and collaborative features.

• Offline Experiments: Run simulations without internet connectivity
• Local SQLite Database: All experiment data stored locally first
• Automatic Sync: PowerSync synchronizes with Supabase when online
• Resilient State: Continue working during network interruptions

• Real-time Help: Ask questions during experiments
• Concept Explanations: AI explains physics, circuits, and astronomy concepts
• Knowledge Retrieval: Tavily integration provides authoritative scientific sources
• Context-Aware: Assistant references your current experiment

• Automatic Analysis: Mastra workflow analyzes completed experiments
• Structured Insights: Generates observations, conclusions, and recommendations
• Learning Progress: AI tracks experiment history for personalized guidance
• Error Detection: Identifies incorrect configurations and suggests fixes

• Automatic Entries: Every experiment creates a structured notebook record
• AI Explanations: Each entry includes AI-generated analysis
• Cross-Device Access: Notebooks sync via PowerSync
• Persistent History: Track learning progress over time

• Real-time Sync: Multiple users can observe/modify same experiment
• Shared State: PowerSync keeps all participants synchronized
• Teacher Dashboard: Educators can monitor student experiments
• Team Learning: Students collaborate on complex simulations

• Frontend: React 18, React Three Fiber, Three.js
• Styling: Tailwind CSS, Framer Motion
• Build Tool: Vite

• Sync Engine: PowerSync (SQLite ↔ Supabase PostgreSQL)
• Backend: Supabase (Authentication, PostgreSQL, RLS)
• AI Framework: Mastra (agents + workflows)
• Knowledge Retrieval: Tavily Search API
• AI Model: OpenAI GPT-4 (via Mastra)

• Node.js 18+
• Supabase account
• PowerSync account
• OpenAI API key
• Tavily API key

npm install

cp .env.example .env

### Database Setup

1. Create a Supabase project
2. Run the schema migration:
```bash
# In Supabase SQL Editor, run:
supabase/schema.sql

3. Configure PowerSync:
   • Create PowerSync instance
   • Connect to your Supabase database
   • Configure sync rules (see docs/powersync-setup.md)

VITE_SUPABASE_URL=your_supabase_project_url
VITE_SUPABASE_ANON_KEY=your_supabase_anon_key
VITE_POWERSYNC_URL=your_powersync_instance_url
VITE_OPENAI_API_KEY=your_openai_api_key
VITE_TAVILY_API_KEY=your_tavily_api_key

npm run dev

The application will be available at http://localhost:3000

User Action → Local SQLite (PowerSync) → Supabase PostgreSQL
                    ↓
              Instant UI Update
                    ↓
         Background Sync (when online)

Experiment Completed → Mastra Workflow Triggered
                            ↓
                    AI Analyzes Data
                            ↓
                  Generates Structured Insights
                            ↓
                Stores in Local DB + Syncs
                            ↓
                  Creates Notebook Entry

User Question → Tavily Search (scientific sources)
                      ↓
              Mastra Agent (GPT-4)
                      ↓
         Contextual Answer + Citations

• Pendulum simulations
• Spring dynamics
• Projectile motion
• Optics experiments
• Real-time measurements with AI analysis

• Drag-and-drop circuit builder
• Component library (batteries, resistors, LEDs, switches)
• Live circuit simulation
• AI error detection for incorrect wiring

• Planetary motion simulation
• Orbital mechanics
• Gravitational interactions
• AI explanations of celestial dynamics