---

⚡ Hosted Version (1 min setup!) (Note: The MCP Client may show some errors, but that's just internal validation failure by the client, so everything will work nonetheless)

Want to start using Earth Engine immediately without any server setup? Use the hosted version through axion-mcp-bridge!

🚀 Quick Start - Zero Configuration

1. Install the bridge globally:

   npm install -g axion-mcp-bridge

2. Add to your MCP client configuration:

   For Claude Desktop (Windows):

   {
     "mcpServers": {
       "axion-mcp": {
         "command": "node",
         "args": [
           "C:\\Users\\YourUsername\\AppData\\Roaming\\npm\\node_modules\\axion-mcp-bridge\\bridge.js"
         ]
       }
     }
   }

   For Mac/Linux:

   {
     "mcpServers": {
       "axion-mcp": {
         "command": "node",
         "args": [
           "/usr/local/lib/node_modules/axion-mcp-bridge/bridge.js"
         ]
       }
     }
   }

   To find your exact path, run: npm root -g and add /axion-mcp-bridge/bridge.js

   That's it! No environment variables, no credentials setup, no server to run. The bridge handles everything.

3. Restart Claude Desktop and start using Earth Engine!

✨ Why Use the Hosted Version?

• 🚀 1-minute setup - Just install and add config
• ☁️ No server management - Everything runs in the cloud
• 🔒 No credentials needed - Secure proxy handles authentication
• ⚡ Always up-to-date - Latest features automatically
• 🌍 Full Earth Engine access - All satellite data and tools available
• 💻 Zero maintenance - No processes to manage

📦 More Details

• NPM Package: axion-mcp-bridge
• Server Status: Always online at https://axion-mcp.onrender.com
• No Google Cloud setup required - We handle the authentication
• Works with any MCP client - Claude Desktop, Cursor, VS Code, etc.

🎯 Try It Now!

Once configured, just ask:

• "Show me vegetation health in California"
• "Create a water map of the Nile River"
• "Analyze urban growth in Tokyo"
• "Monitor deforestation in the Amazon"

---

🎯 The Geospatial AI Revolution

We are witnessing the "iPhone moment" for Earth observation. Just like the iPhone made computing accessible to everyone, Axion Planetary MCP makes petabytes of satellite data accessible through simple conversation.

🔥 The Paradigm Shift

Before: Building geospatial AI required PhD expertise, months of setup, complex APIs, and massive infrastructure.

Now: Anyone can build sophisticated Earth observation AI agents with natural language and one command: npm install

Traditional Path: 1 Expert → 1 Year → 1 Specialized Tool
Our Path:        1 Person → 1 Hour → Unlimited Possibilities

⚡ What Makes This Revolutionary

Axion Planetary MCP is the missing bridge between AI assistants and Earth observation capabilities. It transforms any MCP-compatible client (Claude Desktop, Cline, etc.) into a geospatial intelligence powerhouse with access to Google Earth Engine's massive satellite data catalog.

🌟 What Becomes Possible

👥 Who Can Now Build Geospatial AI Agents:

Before Axion ❌	After Axion ✅
PhD researchers with GEE expertise	Farmers: "Monitor my fields for crop health"
Large corporations with dedicated teams	City Planners: "Track urban expansion patterns"
Government agencies with massive budgets	NGOs: "Monitor deforestation in real-time"
Tech giants with infrastructure	Students: "Study climate change impacts"
	Small Businesses: "Analyze supply chain risks"
	Anyone: Who can install npm and talk to AI

🚀 Real-World Transformations

Precision Agriculture Revolution 🌾

Farmer: "Create an AI agent that monitors my 500-acre farm"
Result: Daily crop health reports, irrigation optimization, 
        pest detection, yield predictions, market timing

Disaster Response at Scale 🔥

Emergency Manager: "Build an agent for wildfire response"
Result: Real-time fire spread prediction, evacuation routing,
        resource allocation, damage assessment, recovery planning

Climate Action Acceleration 🌳

NGO: "Monitor carbon sequestration in our forest projects"
Result: Automated forest health monitoring, carbon calculations,
        impact reporting, donor updates, policy recommendations

🌟 Core Capabilities

Feature	Description
🛫 Satellite Data Access	Direct access to Landsat, Sentinel, MODIS, and 100+ other satellite datasets
📆 30+ Analysis Tools	NDVI, water stress, urban expansion, disaster monitoring, and more
🗺️ Interactive Maps	Generate web-based interactive maps with your analysis results
🤖 5 Pre-trained Models	Wildfire risk, flood prediction, agriculture health, deforestation, water quality
🌾 Smart Crop Classification	ML-powered crop identification with automatic urban/water/vegetation detection
⚡ Real-time Processing	Process live satellite data on-demand
📦 Export Capabilities	Export results as GeoTIFF, create animations, generate reports

🏝️ The Foundation Architecture

🎆 Why This is the Perfect Foundation

We've built the "LEGO blocks" of geospatial AI that anyone can combine:

┌─────────────────────────────────┐
│     Future AI Agents            │
├─────────────────────────────────┤
│  Agriculture AI | Urban Planning│
│  Disaster Mgmt  | Climate Science│
│  Conservation   | Supply Chain  │
└────────────────┬────────────────┘
                 │ MCP Protocol (Standardized)
                 ▼
┌─────────────────────────────────┐
│    Your Foundation Layer        │
│  • Earth Engine Integration    │
│  • Pre-built Models            │
│  • Interactive Visualization   │
│  • Authentication Handling     │
└─────────────────────────────────┘

Core Building Blocks:

• 🛫 Data Access: 100+ satellite datasets
• 🔬 Analysis Tools: NDVI, change detection, classification
• 🗺️ Visualization: Interactive maps, animations
• 🤖 Pre-trained Models: Wildfire, flood, agriculture, deforestation
• 📆 Export Capabilities: GeoTIFF, reports, APIs

🌊 The Network Effect

Once this gains traction, it creates a virtuous cycle:

1. More Users → More use cases discovered
2. More Use Cases → More specialized models needed
3. More Models → More valuable to new users
4. More Value → Attracts more developers
5. Better Tools → Attracts more users

Result: Geospatial AI becomes as common as web development 🌍

---

📋 Prerequisites

Ready to be part of the revolution? Ensure you have:

• ✅ Node.js 18+ installed (Download here)
• ✅ Google Cloud Account (free tier works)
• ✅ MCP-compatible Client (Claude Desktop, Cline, etc.)
• ✅ 4GB RAM minimum (8GB recommended)
• ✅ 2GB free disk space

📦 Local Installation (Self-Hosted Setup)

Want to run your own server locally? Follow these steps to set up the full package:

Option 1: Global Installation (Recommended)

Install globally to use the axion-mcp CLI command from anywhere:

npm install -g axion-planetary-mcp@latest

Or with yarn:

yarn global add axion-planetary-mcp@latest

Option 2: Local Installation

For project-specific installation:

npm install axion-planetary-mcp@latest

Verify Installation

After installation, verify it worked:

# For global installation
axion-mcp --version

# Check where it's installed
npm list -g axion-planetary-mcp

Update to Latest Version

npm update -g axion-planetary-mcp

🔑 Google Earth Engine Setup (REQUIRED)

Step 1: Create Google Cloud Project

1. Go to Google Cloud Console
2. Click "Create Project" or select existing project
3. Give it a name (e.g., "earth-engine-mcp")
4. Note your Project ID - you'll need this

Step 2: Enable Required APIs

In your Google Cloud project, enable these APIs:

1. Go to APIs & Services → Enable APIs and Services
2. Search and enable:
   • ✅ Earth Engine API (CRITICAL!)
   • ✅ Cloud Storage API (for exports)
   • ✅ Cloud Resource Manager API

Step 3: Create Service Account

1. Go to IAM & Admin → Service Accounts
2. Click "+ CREATE SERVICE ACCOUNT"
3. Fill in:
   • Name: earth-engine-sa
   • ID: (auto-generated)
   • Description: "Service account for Earth Engine MCP"
4. Click "CREATE AND CONTINUE"

Step 4: Assign IAM Roles

Add these EXACT roles to your service account:

Role	Why It's Needed
Earth Engine Resource Admin (Beta)	Full access to Earth Engine resources
Earth Engine Resource Viewer (Beta)	Read access to Earth Engine datasets
Service Usage Consumer	Use Google Cloud services
Storage Admin	Manage exports to Cloud Storage
Storage Object Creator	Create export files

How to add roles:

1. In the "Grant this service account access" section
2. Click "Add Role"
3. Search for each role above and add it
4. Click "CONTINUE" then "DONE"

Step 5: Generate JSON Key

1. Click on your newly created service account
2. Go to "Keys" tab
3. Click "ADD KEY" → "Create new key"
4. Choose JSON format
5. Click "CREATE" - file downloads automatically
6. SAVE THIS FILE SECURELY! You'll need it for authentication

Step 6: Register for Earth Engine

1. Go to Earth Engine Sign Up
2. Select "Use with a Cloud Project"
3. Enter your Project ID from Step 1
4. Complete the registration

Step 7: Register Your Service Account with Earth Engine

CRITICAL STEP: Your service account must be registered with Earth Engine to access data!

1. Go to Earth Engine Service Accounts
2. Click "Register a service account"
3. Enter your service account email (format: [email protected])
4. Click "Register"
5. Wait for confirmation (usually instant)

To find your service account email:

• Go to Google Cloud Console
• Navigate to IAM & Admin → Service Accounts
• Copy the email address of your earth-engine-sa account

Step 8: Save Credentials

Save your JSON key file to one of these locations:

Windows:

# Create directory if it doesn't exist
New-Item -ItemType Directory -Force -Path "$env:USERPROFILE\.config\earthengine"

# Copy your key file there
Copy-Item "C:\Downloads\your-key-file.json" "$env:USERPROFILE\.config\earthengine\credentials.json"

Mac/Linux:

# Create directory if it doesn't exist
mkdir -p ~/.config/earthengine

# Copy your key file there
cp ~/Downloads/your-key-file.json ~/.config/earthengine/credentials.json

Alternative: Set environment variable

# Windows
set GOOGLE_APPLICATION_CREDENTIALS=C:\path\to\your\credentials.json

# Mac/Linux
export GOOGLE_APPLICATION_CREDENTIALS=/path/to/your/credentials.json

🚀 Complete Setup Guide

1️⃣ Run Setup Wizard

After installing the package, run:

axion-mcp

This wizard will:

• ✅ Check your Earth Engine credentials
• ✅ Generate MCP configuration
• ✅ Provide exact setup instructions

2️⃣ Start the Next.js Backend (CRITICAL!)

The MCP server requires a Next.js backend to be running.

Open a NEW terminal window and run:

# Navigate to the package directory (path shown by setup wizard)
# Windows example:
cd C:\Users\[YourUsername]\AppData\Roaming\npm\node_modules\axion-planetary-mcp

# Mac example:
cd /usr/local/lib/node_modules/axion-planetary-mcp

# Start the server
npm run start:next

You should see:

▲ Next.js 15.2.4
- Local: http://localhost:3000
✓ Ready

⚠️ IMPORTANT: Keep this terminal window open while using the MCP client!

3️⃣ Configure Your MCP Client

The setup wizard shows you a JSON configuration. Add it to your MCP client's config file:

Claude Desktop Config Locations:

OS	Config File Location
Windows	%APPDATA%\Claude\claude_desktop_config.json
Mac	~/Library/Application Support/Claude/claude_desktop_config.json
Linux	~/.config/claude/claude_desktop_config.json

Example Configuration:

{
  "mcpServers": {
    "axion-planetary": {
      "command": "node",
      "args": ["C:/Users/YourName/.../axion-planetary-mcp/mcp-sse-complete.cjs"],
      "env": {
        "GOOGLE_APPLICATION_CREDENTIALS": "C:/Users/YourName/.config/earthengine/credentials.json"
      }
    }
  }
}

4️⃣ Restart Your MCP Client

Completely quit and restart your MCP client to load the new configuration.

5️⃣ Test It!

Ask your MCP client:

• "Show me current NDVI for California farmland"
• "Create a crop classification map for Iowa"
• "Analyze urban heat islands in Los Angeles"

✨ Features

🛠️ Core Tools

1. Data Discovery & Access (earth_engine_data)

• Search satellite datasets
• Filter by date, location, cloud cover
• Access dataset metadata
• Get region boundaries

2. Processing & Analysis (earth_engine_process)

• Calculate vegetation indices (NDVI, EVI, SAVI, etc.)
• Create cloud-free composites
• Perform terrain analysis
• Generate statistics and time series

3. Export & Visualization (earth_engine_export)

• Export to GeoTIFF format
• Generate thumbnails
• Create map tiles
• Track export status

4. Interactive Maps (earth_engine_map)

• Create web-based interactive maps
• Visualize large regions
• Multiple layer support
• Share results via URL

5. System Operations (earth_engine_system)

• Check authentication status
• Execute custom Earth Engine code
• Monitor system health

🤖 Pre-trained Models

Model	Use Case	Example
🔥 Wildfire Risk	Assess fire danger zones	"Analyze wildfire risk in California"
💧 Flood Prediction	Identify flood-prone areas	"Show flood risk for Houston"
🌾 Agriculture Health	Monitor crop conditions	"Check crop health in Iowa farmland"
🌲 Deforestation	Detect forest loss	"Monitor Amazon deforestation since 2020"
🏊 Water Quality	Analyze water bodies	"Assess water quality in Lake Tahoe"

🌾 Advanced Crop Classification

The crop classification tool includes:

• Automatic augmentation with urban, water, and vegetation classes
• Pre-configured training data for major US states
• Multiple classifiers: Random Forest, SVM, CART, Naive Bayes
• Interactive result maps

Supported regions with built-in training data:

• Iowa (corn, soybean)
• California (almonds, grapes, citrus, rice)
• Texas (cotton, wheat, sorghum)
• Kansas (wheat, corn, sorghum, soybean)
• Nebraska (corn, soybean, wheat)
• Illinois (corn, soybean, wheat)

📚 The Magic: Natural Language → Earth Intelligence

Just talk to your AI assistant like you would a geospatial expert:

🌾 Agriculture & Food Security

"How healthy are the crops in Iowa this season?"

"Which fields in Nebraska need irrigation most urgently?"

"Create a crop classification map showing corn vs soybean distribution"

"Predict wheat yields for Kansas based on current conditions"

🔥 Disaster Response & Climate

"Show me wildfire risk zones in California with evacuation routes"

"Track the flood extent after Hurricane Ian in real-time"

"Which areas of Texas are most vulnerable to drought?"

"Monitor deforestation in the Amazon and calculate carbon impact"

🏢 Urban Planning & Development

"How fast is Phoenix expanding and where should we plan infrastructure?"

"Identify urban heat islands in New York City for cooling strategies"

"Track construction progress in Austin's development zones"

"Analyze land use changes in Seattle over the past 5 years"

💧 Water Resources & Environment

"How are Lake Mead's water levels changing over time?"

"Detect harmful algae blooms in the Great Lakes system"

"Monitor coastal erosion patterns in Miami Beach"

"Assess water quality in Lake Tahoe using satellite data"

🌍 Conservation & Research

"Create a time-lapse animation of Las Vegas urban growth since 2000"

"Export detailed NDVI analysis for my research area as GeoTIFF"

"Generate false color imagery highlighting vegetation stress patterns"

"Calculate forest carbon sequestration in protected areas"

✨ The Result: Instant expert-level geospatial analysis with interactive maps, detailed reports, and actionable insights.

---

🚀 Ready to Build the Future?

Every revolution starts with early adopters. The farmers who first used tractors. The businesses that first went online. The developers who first embraced cloud computing.

Now it's your turn to be part of the geospatial AI revolution.

🌟 Why Start Now?

• ⏰ Perfect Timing: AI + Earth observation converging at exactly the right moment
• 🌍 Urgent Need: Climate change, food security, and disasters require immediate action
• 📈 First-Mover Advantage: Build expertise while the field is still emerging
• 🤝 Growing Community: Join thousands already exploring new possibilities
• ✅ Proven Foundation: Built on Google Earth Engine's enterprise-grade infrastructure

The question isn't whether geospatial AI will transform every industry—it's whether you'll be leading that transformation or watching from the sidelines.

---

🎓 Technical Architecture (For the Curious)

┌─────────────────┐
│   MCP Client    │  (Claude Desktop, Cline, etc.)
└────────┬────────┘
         │ stdio/JSON-RPC
         ▼
┌─────────────────┐
│  MCP SSE Bridge │  (mcp-sse-complete.cjs)
└────────┬────────┘
         │ HTTP/SSE
         ▼
┌─────────────────┐
│  Next.js API    │  (localhost:3000/api/mcp/sse)
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  Earth Engine   │  (Processing & Analysis)
└─────────────────┘

The system uses a bridge architecture where:

1. MCP client communicates via stdio/JSON-RPC
2. Bridge converts to HTTP/Server-Sent Events
3. Next.js backend handles Earth Engine operations
4. Results flow back through the same pipeline

🔧 Troubleshooting

"MCP server not responding"

Solution:

1. ✅ Ensure Next.js server is running in separate terminal
2. ✅ Check http://localhost:3000 is accessible
3. ✅ Restart your MCP client
4. ✅ Verify config file path uses forward slashes (/)

"Earth Engine authentication failed"

Solution:

1. ✅ Verify credentials.json exists and is valid JSON
2. ✅ Confirm all 5 IAM roles are assigned to service account
3. ✅ Check Earth Engine API is enabled in Google Cloud
4. ✅ Ensure you've registered for Earth Engine with your project
5. ✅ CRITICAL: Verify service account is registered at https://code.earthengine.google.com/register

"Request failed" errors

Solution:

1. ✅ Next.js server MUST be running (npm run start:next)
2. ✅ Port 3000 must be free
3. ✅ Check Windows Firewall isn't blocking port 3000

Maps not displaying

Solution:

1. ✅ Explicitly request map creation: "create a map showing..."
2. ✅ Visit http://localhost:3000 to verify server is running
3. ✅ Check browser console for errors

Port 3000 already in use

Solution:

# Use different port
$env:PORT=3001; npm run start:next  # Windows
PORT=3001 npm run start:next         # Mac/Linux

Installation issues

Solution:

1. ✅ Use Node.js 18 or higher: node --version
2. ✅ Clear npm cache: npm cache clean --force
3. ✅ Run as Administrator (Windows)
4. ✅ Try without -g: npm install axion-planetary-mcp

🌟 Pro Tips

Optimize Performance

• Use scale parameter for faster processing (higher number = lower resolution)
• Filter by cloud cover for cleaner imagery
• Specify date ranges to limit data processing

Better Results

• Request "cloud-free composite" for clearer images
• Use "median composite" to reduce noise
• Add "with interactive map" to get visual results

Advanced Features

• Chain operations: "Calculate NDVI, then create a map"
• Export results: "Export the analysis as GeoTIFF"
• Time series: "Show monthly changes over 2024"

📊 Available Datasets

Popular datasets you can access:

Dataset	Description	Best For
Sentinel-2	10m resolution, 5-day revisit	Detailed land analysis
Landsat 8/9	30m resolution, 16-day revisit	Long-term monitoring
MODIS	Daily imagery, 250m-1km resolution	Large area analysis
Sentinel-1	Radar imagery, works through clouds	Flood detection
NAIP	1m resolution aerial imagery (US only)	High-detail mapping

📈 Performance & Limits

• Processing Scale: 10m to 1000m resolution
• Region Size: Best for areas under 10,000 km²
• Time Range: Data from 1972 to present
• Export Size: Up to 10GB per file
• Rate Limits: Respects Earth Engine quotas

🤝 Contributing

We welcome contributions! Please feel free to:

• Report bugs via GitHub Issues
• Submit pull requests
• Suggest new features
• Improve documentation

📄 License

MIT License - feel free to use in your projects!

💬 Support

• GitHub Issues: Report bugs or request features
• Discussions: Ask questions and share tips
• Documentation: Wiki and guides

🙏 Acknowledgments

• Google Earth Engine team for the amazing platform
• Anthropic for the MCP protocol
• The open-source geospatial community
• All contributors and users

---