Cursor / Claude Code / Any MCP Client
        │
        ▼
┌─────────────────────────────────────────┐
│         Wave OS MCP Server              │
│                                         │
│  Layer 1: Base44 Backend (Data)        │
│  ├─ Entity CRUD (create, read, update)  │
│  ├─ Aggregation pipelines               │
│  ├─ Backend function calls              │
│  └─ File uploads                        │
│                                         │
│  Layer 2: Theta Compute (GPU + Chain)   │
│  ├─ AI model inference                  │
│  ├─ GPU instance management             │
│  ├─ Smart contract deployment           │
│  ├─ Contract reads (view functions)     │
│  ├─ Wallet balance queries              │
│  └─ Transaction lookups                 │
│                                         │
│  Layer 3: Wave OS Intelligence          │
│  ├─ Morning briefing (Chief of Staff)   │
│  ├─ Triage scanning                     │
│  ├─ Follow-up tracking                  │
│  ├─ Meeting prep                        │
│  ├─ Memory save/recall                  │
│  ├─ Sub-agent delegation                │
│  └─ Wave Assistant chat                 │
└─────────────────────────────────────────┘
        │              │              │
        ▼              ▼              ▼
   Base44 API    Theta EdgeCloud   Theta RPC
   (database,    (GPU compute,     (blockchain,
    functions,    AI inference)     smart contracts)
    auth, files)

cd wave-mcp-server
npm install

export BASE44_APP_ID="your_app_id"
export BASE44_API_KEY="your_api_key"
export THETA_API_KEY="your_theta_api_key"
export THETA_PROJECT_ID="your_theta_project_id"

npm run build

{
  "mcpServers": {
    "wave-os": {
      "command": "node",
      "args": ["/path/to/wave-mcp-server/dist/index.js"],
      "env": {
        "BASE44_APP_ID": "your_app_id",
        "BASE44_API_KEY": "your_api_key",
        "THETA_API_KEY": "your_theta_api_key",
        "THETA_PROJECT_ID": "your_theta_project_id"
      }
    }
  }
}

You: "Create a Customer entity on my Base44 app"
AI: [calls create_entity] → Entity created with fields: name, email, status

You: "Add 5 sample customers"
AI: [calls create_records] → 5 records inserted

You: "Run sentiment analysis on these customers"
AI: [calls call_function to deploy sentiment function]
    [calls theta_run_inference to analyze each customer]
    [calls update_records to store sentiment scores]

You: "Show me the results"
AI: [calls read_records with fields projection] → Returns enriched data

You: "What needs my attention today?"
AI: [calls wave_triage] → Returns: 2 calendar conflicts, 1 credit warning, 3 overdue tasks

You: "Save a note that Customer X is VIP"
AI: [calls wave_save_memory] → Memory saved with category "contact"

You: "Delegate a follow-up task to a sub-agent for Customer X"
AI: [calls wave_delegate_subagent] → Sub-agent spawned, task assigned

You: "Check the treasury wallet balance"
AI: [calls theta_get_balance] → Returns TFUEL and WAVE balances

You: "How many token holders do we have?"
AI: [calls read_records on TokenHolder entity] → Returns holder list

You: "Deploy a new TNT20 token contract"
AI: [calls theta_deploy_contract] → Contract deployed to mainnet
    [calls create_records to log the deployment]