WakeIQX

Appearance

Getting Started with Wake Intelligence ​

Welcome to WakeIQX - the temporal intelligence MCP server that gives your AI agent a brain with memory of the Past, awareness of the Present, and vision of the Future.

Prerequisites ​

Before you begin, ensure you have:

  • Node.js 20.x or higher - Download here
  • Cloudflare account - Sign up free
  • Wrangler CLI - Install with npm install -g wrangler
  • Claude Desktop (optional) - For MCP integration

Installation ​

1. Clone the Repository ​

bash
git clone https://github.com/semanticintent/semantic-wake-intelligence-mcp.git
cd semantic-wake-intelligence-mcp

2. Install Dependencies ​

bash
npm install

3. Create D1 Database ​

bash
# Create the database
wrangler d1 create wake-intelligence

# This will output your database ID - copy it!

4. Configure Wrangler ​

Update wrangler.jsonc with your database ID:

jsonc
{
  "d1_databases": [{
    "binding": "DB",
    "database_name": "wake-intelligence",
    "database_id": "YOUR-DATABASE-ID-HERE"  // ← Paste your ID here
  }]
}

5. Run Database Migrations ​

Wake Intelligence requires 5 migrations (one per layer + initial schema):

bash
# Migration 0001: Core context table
wrangler d1 execute wake-intelligence --local --file=migrations/0001_initial_schema.sql

# Migration 0002: Layer 1 - Causality Engine (Past)
wrangler d1 execute wake-intelligence --local --file=migrations/0002_add_causality_engine.sql

# Migration 0003: Layer 2 - Memory Manager (Present)
wrangler d1 execute wake-intelligence --local --file=migrations/0003_add_memory_manager.sql

# Migration 0004: Layer 3 - Propagation Engine (Future)
wrangler d1 execute wake-intelligence --local --file=migrations/0004_add_propagation_engine.sql

# Migration 0005: Layer 4 - Meta-Learning (Adaptive)
wrangler d1 execute wake-intelligence --local --file=migrations/0005_add_meta_learning.sql

For production deployment, remove --local:

bash
wrangler d1 execute wake-intelligence --file=migrations/0001_initial_schema.sql
wrangler d1 execute wake-intelligence --file=migrations/0002_add_causality_engine.sql
wrangler d1 execute wake-intelligence --file=migrations/0003_add_memory_manager.sql
wrangler d1 execute wake-intelligence --file=migrations/0004_add_propagation_engine.sql
wrangler d1 execute wake-intelligence --file=migrations/0005_add_meta_learning.sql

6. Start Development Server ​

bash
npm run dev

Your MCP server is now running at http://localhost:8787 🎉

Verify Installation ​

Test with curl ​

bash
# Health check
curl http://localhost:8787/

# List available tools
curl -X POST http://localhost:8787/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'

You should see a list of 19 MCP tools including:

  • save_context, load_context, search_context
  • reconstruct_reasoning, build_causal_chain, get_causality_stats
  • get_memory_stats, get_memory_health
  • update_predictions, get_high_value_contexts, get_propagation_stats
  • get_learning_stats, reindex_project, admin_reindex_all
  • get_causal_graph, ingest_rune_manifest
  • And more — full list

Deploy to Production ​

When you're ready to deploy to Cloudflare Workers:

bash
# Deploy to production
npm run deploy

Your MCP server will be available at:

https://semantic-wake-intelligence-mcp.YOUR-ACCOUNT.workers.dev

Connect to Claude Desktop ​

To use Wake Intelligence with Claude Desktop:

1. Install mcp-remote ​

bash
npm install -g mcp-remote

2. Configure Claude Desktop ​

Edit your Claude Desktop configuration file:

macOS/Linux: ~/Library/Application Support/Claude/claude_desktop_config.jsonWindows: %APPDATA%\Claude\claude_desktop_config.json

Add this configuration:

json
{
  "mcpServers": {
    "wake-intelligence": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://localhost:8787/sse"
      ]
    }
  }
}

For production deployment (with Bearer token auth):

json
{
  "mcpServers": {
    "wake-intelligence": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://YOUR-CUSTOM-DOMAIN.dev/sse",
        "--header",
        "Authorization: Bearer YOUR_MCP_SECRET"
      ]
    }
  }
}

Tip: Use a custom domain (wake.yourdomain.dev) rather than the raw workers.dev URL — it masks your CF account identifier and lets you add WAF rate limiting rules scoped to that domain.

3. Restart Claude Desktop ​

Close and reopen Claude Desktop. You should see Wake Intelligence tools available in the tool palette.

First Steps with Wake Intelligence ​

Save Your First Context ​

Claude, save this context:
"Started Wake Intelligence setup. Database created and migrations successful. Ready to build temporal intelligence features."

Use action type: implementation
Project: wake-intelligence-setup

Load Contexts ​

Claude, load contexts for project "wake-intelligence-setup"

Reconstruct Reasoning ​

Claude, reconstruct the reasoning for context [context-id]

Check Memory Stats ​

Claude, show me memory statistics for project "wake-intelligence-setup"

Understanding the 3 Layers ​

Now that Wake Intelligence is running, let's understand what each layer does:

Layer 1: Causality Engine (Past - WHY) ​

Automatically tracks WHY contexts were created:

typescript
// Every save can include causality metadata
{
  actionType: "decision",  // What kind of action
  rationale: "Chose Redis for caching because...",  // WHY
  dependencies: ["context-123", "context-456"]  // Related contexts
}

Use it to:

  • Understand decision history
  • Build causal chains
  • Reconstruct reasoning

Layer 2: Memory Manager (Present - HOW) ​

Automatically classifies contexts by relevance:

  • ACTIVE - Accessed < 1 hour ago
  • RECENT - Accessed 1-24 hours ago
  • ARCHIVED - Accessed 1-30 days ago
  • EXPIRED - Accessed > 30 days ago

Use it to:

  • Prioritize recent contexts
  • Automatically prune old data
  • Track access patterns

Layer 3: Propagation Engine (Future - WHAT) ​

Predicts what contexts you'll need next:

typescript
// Composite prediction score
score = temporal_weight * temporal + causal_weight * causal + frequency_weight * frequency

// Reasons: ["recently_accessed", "causal_chain_root", "high_frequency"]

Use it to:

  • Pre-fetch likely-needed contexts
  • Optimize cache strategy
  • Identify patterns

Layer 4: Meta-Learning (Adaptive - HOW WELL) ​

Observes prediction accuracy and tunes Layer 3 weights per project:

typescript
// After 20+ recorded outcomes, Layer 4 adjusts weights
// e.g., if temporal signal was consistently right → temporal_weight increases
get_learning_stats({ project: "api-service" });
// → Temporal weight: 0.52 (learned), Causal: 0.28, Frequency: 0.20

Use it to:

  • Understand which signal drives your project's predictions
  • Verify learning has activated (requires ≥20 outcomes)
  • Monitor weight drift over time

Authentication ​

All requests to the deployed Wake Intelligence server require a Bearer token:

bash
curl -X POST https://your-worker.workers.dev/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_MCP_SECRET" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'

Set MCP_SECRET in your Cloudflare Worker environment (via wrangler secret put MCP_SECRET). When connecting via Claude Desktop, add the token to your mcp-remote args:

json
{
  "mcpServers": {
    "wake-intelligence": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://your-worker.workers.dev/sse",
        "--header",
        "Authorization: Bearer YOUR_MCP_SECRET"
      ]
    }
  }
}

Next Steps ​

Now that you have Wake Intelligence running:

  1. Explore the Architecture - 3-Layer Brain Architecture
  2. Learn the Tools - MCP Tools Reference
  3. Dive into APIs - API Documentation
  4. Understand the Pattern - Semantic Intent

Troubleshooting ​

Database Connection Issues ​

Problem: Error: D1 database not found

Solution: Ensure your wrangler.jsonc has the correct database ID:

bash
# List your databases
wrangler d1 list

# Verify the ID matches your configuration

Migration Errors ​

Problem: Error: table already exists

Solution: Migrations are idempotent. If a table exists, the migration will skip creation. Check which migrations have run:

bash
wrangler d1 execute wake-intelligence --local --command="SELECT name FROM sqlite_master WHERE type='table'"

Local Development Port Conflicts ​

Problem: Error: Address already in use (port 8787)

Solution: Either stop the other process using port 8787, or change the port in wrangler.jsonc:

jsonc
{
  "dev": {
    "port": 8788  // Use a different port
  }
}

Claude Desktop Not Seeing Tools ​

Problem: Tools don't appear in Claude Desktop

Solutions:

  1. Verify your claude_desktop_config.json path is correct
  2. Restart Claude Desktop completely (quit and reopen)
  3. Check that the MCP server is running: curl http://localhost:8787/
  4. Verify mcp-remote is installed: npm list -g mcp-remote

Getting Help ​

What's Next? ​

You're now ready to use Wake Intelligence! Here are some suggested next steps:

Welcome to temporal intelligence! 🧠

The wake persists. 🐦