Integrates with Google Cloud Platform for service account authentication, API management, and Cloud Storage operations to support Earth Engine data exports and processing workflows.
Provides comprehensive access to Google Earth Engine's satellite data catalog and geospatial analysis capabilities, enabling petabyte-scale Earth observation, vegetation analysis, crop classification, disaster monitoring, and interactive map generation with 30+ analysis tools and pre-trained ML models.
π Axion Planetary MCP
The Foundation for Democratizing Geospatial AI Agents
π Making Earth Observation as Easy as Having a Conversation
From PhD-level complexity to natural language queries in one install
"Show me crop health in Iowa" β’ "Analyze wildfire risk in California" β’ "Track deforestation in Amazon"
π― The Revolution β’ β‘ Quick Start β’ π What's Possible β’ π οΈ Setup
β‘ 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!
π Quick Start - Zero Configuration
Install the bridge globally:
npm install -g axion-mcp-bridgeAdd 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.
Restart Claude Desktop and start using Earth Engine!
β¨ Why Use the Hosted Version?
π 1-minute setup - Just install and add config
βοΈ No server management - Everything runs in the cloud
π No credentials needed - Secure proxy handles authentication
β‘ Always up-to-date - Latest features automatically
π Full Earth Engine access - All satellite data and tools available
π» Zero maintenance - No processes to manage
π¦ 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.
π― 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"
π― 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.
π₯ 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
β‘ 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.
π What Becomes Possible
π₯ Who Can Now Build Geospatial AI Agents:
Before Axion β | After Axion β |
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 |
π Real-World Transformations
Precision Agriculture Revolution πΎ
Disaster Response at Scale π₯
Climate Action Acceleration π³
π Core Capabilities
Feature | Description |
π« Satellite Data Access | Direct access to Landsat, Sentinel, MODIS, and 100+ other satellite datasets |
π 30+ Analysis Tools | NDVI, water stress, urban expansion, disaster monitoring, and more |
πΊοΈ Interactive Maps | Generate web-based interactive maps with your analysis results |
π€ 5 Pre-trained Models | Wildfire risk, flood prediction, agriculture health, deforestation, water quality |
πΎ Smart Crop Classification | ML-powered crop identification with automatic urban/water/vegetation detection |
β‘ Real-time Processing | Process live satellite data on-demand |
π¦ Export Capabilities | Export results as GeoTIFF, create animations, generate reports |
ποΈ The Foundation Architecture
π Why This is the Perfect Foundation
We've built the "LEGO blocks" of geospatial AI that anyone can combine:
Core Building Blocks:
π« Data Access: 100+ satellite datasets
π¬ Analysis Tools: NDVI, change detection, classification
πΊοΈ Visualization: Interactive maps, animations
π€ Pre-trained Models: Wildfire, flood, agriculture, deforestation
π Export Capabilities: GeoTIFF, reports, APIs
π The Network Effect
Once this gains traction, it creates a virtuous cycle:
More Users β More use cases discovered
More Use Cases β More specialized models needed
More Models β More valuable to new users
More Value β Attracts more developers
Better Tools β Attracts more users
Result: Geospatial AI becomes as common as web development π
π Prerequisites
Ready to be part of the revolution? Ensure you have:
β Node.js 18+ installed (Download here)
β Google Cloud Account (free tier works)
β MCP-compatible Client (Claude Desktop, Cline, etc.)
β 4GB RAM minimum (8GB recommended)
β 2GB free disk space
π¦ 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:
Or with yarn:
Option 2: Local Installation
For project-specific installation:
Verify Installation
After installation, verify it worked:
Update to Latest Version
π Google Earth Engine Setup (REQUIRED)
Step 1: Create Google Cloud Project
Go to Google Cloud Console
Click "Create Project" or select existing project
Give it a name (e.g., "earth-engine-mcp")
Note your Project ID - you'll need this
Step 2: Enable Required APIs
In your Google Cloud project, enable these APIs:
Go to APIs & Services β Enable APIs and Services
Search and enable:
β Earth Engine API (CRITICAL!)
β Cloud Storage API (for exports)
β Cloud Resource Manager API
Step 3: Create Service Account
Go to IAM & Admin β Service Accounts
Click "+ CREATE SERVICE ACCOUNT"
Fill in:
Name:
earth-engine-sa
ID: (auto-generated)
Description: "Service account for Earth Engine MCP"
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:
In the "Grant this service account access" section
Click "Add Role"
Search for each role above and add it
Click "CONTINUE" then "DONE"
Step 5: Generate JSON Key
Click on your newly created service account
Go to "Keys" tab
Click "ADD KEY" β "Create new key"
Choose JSON format
Click "CREATE" - file downloads automatically
SAVE THIS FILE SECURELY! You'll need it for authentication
Step 6: Register for Earth Engine
Go to Earth Engine Sign Up
Select "Use with a Cloud Project"
Enter your Project ID from Step 1
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!
Click "Register a service account"
Enter your service account email (format:
earth-engine-sa@YOUR-PROJECT-ID.iam.gserviceaccount.com
)Click "Register"
Wait for confirmation (usually instant)
To find your service account email:
Go to Google Cloud Console
Navigate to IAM & Admin β 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:
Mac/Linux:
Alternative: Set environment variable
π Complete Setup Guide
1οΈβ£ Run Setup Wizard
After installing the package, run:
This wizard will:
β Check your Earth Engine credentials
β Generate MCP configuration
β Provide exact setup instructions
2οΈβ£ Start the Next.js Backend (CRITICAL!)
The MCP server requires a Next.js backend to be running.
Open a NEW terminal window and run:
You should see:
β οΈ IMPORTANT: Keep this terminal window open while using the MCP client!
3οΈβ£ 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 |
|
Mac |
|
Linux |
|
Example Configuration:
4οΈβ£ Restart Your MCP Client
Completely quit and restart your MCP client to load the new configuration.
5οΈβ£ 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"
β¨ Features
π οΈ 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
π€ Pre-trained Models
Model | Use Case | Example |
π₯ Wildfire Risk | Assess fire danger zones | "Analyze wildfire risk in California" |
π§ Flood Prediction | Identify flood-prone areas | "Show flood risk for Houston" |
πΎ Agriculture Health | Monitor crop conditions | "Check crop health in Iowa farmland" |
π² Deforestation | Detect forest loss | "Monitor Amazon deforestation since 2020" |
π Water Quality | Analyze water bodies | "Assess water quality in Lake Tahoe" |
πΎ 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)
π The Magic: Natural Language β Earth Intelligence
Just talk to your AI assistant like you would a geospatial expert:
πΎ 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"
π₯ 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"
π’ 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"
π§ 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"
π 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"
β¨ The Result: Instant expert-level geospatial analysis with interactive maps, detailed reports, and actionable insights.
π 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.
π Why Start Now?
β° Perfect Timing: AI + Earth observation converging at exactly the right moment
π Urgent Need: Climate change, food security, and disasters require immediate action
π First-Mover Advantage: Build expertise while the field is still emerging
π€ Growing Community: Join thousands already exploring new possibilities
β Proven Foundation: Built on Google Earth Engine's enterprise-grade infrastructure
The question isn't whether geospatial AI will transform every industryβit's whether you'll be leading that transformation or watching from the sidelines.
π Technical Architecture (For the Curious)
The system uses a bridge architecture where:
MCP client communicates via stdio/JSON-RPC
Bridge converts to HTTP/Server-Sent Events
Next.js backend handles Earth Engine operations
Results flow back through the same pipeline
π§ Troubleshooting
"MCP server not responding"
Solution:
β Ensure Next.js server is running in separate terminal
β Check http://localhost:3000 is accessible
β Restart your MCP client
β Verify config file path uses forward slashes (/)
"Earth Engine authentication failed"
Solution:
β Verify credentials.json exists and is valid JSON
β Confirm all 5 IAM roles are assigned to service account
β Check Earth Engine API is enabled in Google Cloud
β Ensure you've registered for Earth Engine with your project
β CRITICAL: Verify service account is registered at https://code.earthengine.google.com/register
"Request failed" errors
Solution:
β Next.js server MUST be running (npm run start:next)
β Port 3000 must be free
β Check Windows Firewall isn't blocking port 3000
Maps not displaying
Solution:
β Explicitly request map creation: "create a map showing..."
β Visit http://localhost:3000 to verify server is running
β Check browser console for errors
Port 3000 already in use
Solution:
Installation issues
Solution:
β Use Node.js 18 or higher:
node --version
β Clear npm cache:
npm cache clean --force
β Run as Administrator (Windows)
β Try without
-g
:npm install axion-planetary-mcp
π 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"
π 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 |
π Performance & Limits
Processing Scale: 10m to 1000m resolution
Region Size: Best for areas under 10,000 kmΒ²
Time Range: Data from 1972 to present
Export Size: Up to 10GB per file
Rate Limits: Respects Earth Engine quotas
π€ Contributing
We welcome contributions! Please feel free to:
Report bugs via GitHub Issues
Submit pull requests
Suggest new features
Improve documentation
π License
MIT License - feel free to use in your projects!
π¬ Support
GitHub Issues: Report bugs or request features
Discussions: Ask questions and share tips
Documentation: Wiki and guides
π Acknowledgments
Google Earth Engine team for the amazing platform
Anthropic for the MCP protocol
The open-source geospatial community
All contributors and users
π The Future is Now
This isn't just a toolβit's the foundation of a revolution.
We're democratizing Earth observation, making geospatial intelligence as accessible as sending a text message.
Join the thousands already building the future of geospatial AI.
π What Will You Build?
πΎ Agricultural AI that saves crops? β’ π₯ Wildfire prediction that saves lives? β’ π³ Forest monitoring that fights climate change?
The Earth is waiting. The tools are ready. The only question is: what will you discover?
From PhD-level complexity to conversational simplicity in one command β¨
Built with β€οΈ to accelerate humanity's response to our biggest challenges
This server cannot be installed
local-only server
The server can only run on the client's local machine because it depends on local resources.
Enables MCP clients to access Google Earth Engine's satellite imagery and geospatial analysis capabilities. Provides tools for vegetation analysis, crop classification, disaster monitoring, and interactive map creation using petabytes of satellite data.
- The Foundation for Democratizing Geospatial AI Agents
- β‘ 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)
- π― The Geospatial AI Revolution
- π What Becomes Possible
- ποΈ The Foundation Architecture
- π Prerequisites
- π¦ Local Installation (Self-Hosted Setup)
- π Google Earth Engine Setup (REQUIRED)
- π Complete Setup Guide
- β¨ Features
- π The Magic: Natural Language β Earth Intelligence
- π Ready to Build the Future?
- π Technical Architecture (For the Curious)
- π§ Troubleshooting
- π Pro Tips
- π Available Datasets
- π Performance & Limits
- π€ Contributing
- π License
- π¬ Support
- π Acknowledgments
- π The Future is Now