Sentinel Hub MCP for AI. Analyze satellite data and derive spectral indices.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Connect to your AI in seconds.
Sentinel Hub connects your AI client directly to massive Earth observation datasets from Sentinel, Landsat, and other missions. Use it to search global satellite imagery by location, calculate specialized indices like NDVI or burn severity, and process the resulting data into actionable maps—all without downloading terabytes of raw files.
What your AI can do
Catalog search
Finds available satellite images by specifying the data collection, a bounding box, and a date range.
Check sentinel hub status
Verifies if your API credentials are working correctly and confirms connectivity to the service.
Generate false color evalscript
Creates a processing script that emphasizes vegetation (red), urban areas (cyan/grey), and water (dark blue).
Find available imagery metadata by specifying a collection, bounding box, and date range.
Create ready-to-use evaluation scripts (evalscripts) to colorize specific bands or calculate standard indices like NDVI.
Run the generated evalscript against a specified geographic area and date range, outputting processed satellite data.
Derive statistical metrics (mean, max, std dev) for an area of interest across daily, weekly, or monthly time series.
Search only for satellite scenes that fall below a specified cloud cover threshold.
Ask an AI about this
Waiting for input…
Sentinel Hub MCP Server: 14 Tools for Geospatial Analysis
These tools let your agent manage the entire geospatial workflow—from finding clean imagery to generating complex statistical reports.
Make your AI actually useful.
Add this MCP to Claude, Cursor, or Windsurf and your AI stops guessing. It gets real tools to look things up, take action, and handle the stuff you keep doing by hand.
Start using Sentinel Hub on VinkiusCatalog Search
Finds available satellite images by specifying the data collection, a bounding box, and a date range.
Check Sentinel Hub Status
Verifies if your API credentials are working correctly and confirms connectivity to...
Generate False Color Evalscript
Creates a processing script that emphasizes vegetation (red), urban areas...
Generate Ndvi Evalscript
Generates a ready-to-use script for NDVI, coloring the output from dark (water) to...
Generate True Color Evalscript
Creates a script that generates standard RGB color images, making them look like...
Get Catalog Collection
Retrieves detailed information about a specific satellite data collection ID.
Get Catalog Item
Gets deep metadata for one single catalog item, using the ID found during a search.
Get Statistics
Calculates statistical summaries (mean, min, max) over an area of interest across...
Get User Info
Checks your account credentials and reports on available service quotas.
List Band Combinations
Lists all predefined spectral indices and band combinations, including True Color...
List Catalog Collections
Provides a list of every major satellite data source available, such as Sentinel-2...
Process Image
Executes the full analysis: processes imagery using your custom script over a defined area and date range.
Search By Tile
Searches for all available satellite scenes within a specific geographical MGRS tile identifier.
Search Cloud Free
Filters the search results to only include imagery that meets or exceeds your...
Security and governance baked right in.
Pick your AI client below to get set up. Just create a Vinkius account, subscribe, and you're instantly up and running. We handle the entire backend infrastructure, delivering out-of-the-box support for HTTPS Streamable, SSE, and OAuth2—zero messy routing required.
Claude AI
Open Claude Settings
Go to claude.ai, click your profile icon, then navigate to Customize → Connectors.
Add Custom Connector
Click the "+" button and select Add custom connector. Paste your Vinkius endpoint URL:
https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp
Replace [YOUR_TOKEN_HERE] with your token
from cloud.vinkius.com. For OAuth-protected servers, expand
Advanced settings to add credentials.
Start a conversation
Open a new chat. The Sentinel Hub integration is available immediately — no restart needed.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Sentinel Hub, then connect any of our 5,000+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 5,000+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Sentinel Hub. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This connection provides 14 powerful capabilities that interface natively with Claude, ChatGPT, Cursor, and other compatible AI platforms. No middleware. No custom integration required.
Analyzing global data used to be an exercise in manual API calls and massive file transfers.
Before dedicated tools like Sentinel Hub, running a basic land-use comparison meant logging into multiple vendor portals. You'd manually search by bounding box, download raw TIF files measured in gigabytes, and then spend hours in GIS software just to stitch together the time series and calculate simple averages.
With this MCP Server, you define the goal—say, comparing vegetation density over two years. Your agent handles the API calls, filters out cloudy images using `search_cloud_free`, generates the NDVI script via `generate_ndvi_evalscript`, and runs the whole process to give you a clean, comparative map. The sheer volume of manual work disappears.
Sentinel Hub MCP Server: Process Imagery with a Custom Evalscript
Previously, generating an index like the Normalized Difference Vegetation Index (NDVI) required deep knowledge of band combinations and manual scripting within specialized software. You had to know which bands were needed—Near-Infrared (NIR) and Red—and how to mathematically combine them.
Now, you tell your agent 'I need NDVI.' The server handles the complex logic: it generates the `generate_ndvi_evalscript` for you, ensuring the correct band inputs are used. You just pass that script into `process_image`, and the resulting data is ready to analyze.
What your AI can actually do with this
You connect your agent straight into massive Earth observation datasets from Sentinel, Landsat, and other missions. You don’t have to manually query dozens of APIs or deal with moving terabytes of raw files; you just tell your client what analysis you need—an index map, a cloud-free view of an area, statistical trends over months—and it runs the whole thing.
Getting Started and Finding Data
You'll first check connectivity by running check_sentinel_hub_status to verify your API credentials are working. You can confirm your account status and available service quotas using get_user_info. To see what data is out there, you use list_catalog_collections to pull a list of every major satellite source, like Sentinel-2 or Landsat. If you need specifics on one of those collections, run list_catalog_collections to get the detailed information for that specific ID via get_catalog_collection.
Searching is all about filtering down massive catalogs. You use catalog_search when you know the general area, specifying a bounding box and date range along with the data collection type. If you're working with a precise geographic identifier, you can narrow your search using search_by_tile based on an MGRS tile ID.
To ensure quality results, you filter out unusable images by running search_cloud_free, which guarantees that any returned imagery meets or exceeds your specified cloud cover threshold. Once you've found a scene, you get deep metadata for one specific item using get_catalog_item.
Building the Analysis Logic
To figure out what kind of analysis to run, you start by listing available indices. The list_band_combinations tool shows all predefined spectral indexes and band pairings, including True Color, NDVI, and Burn Severity (NBR). You'll use this list to create your custom scripts. To make it look like a natural photograph, you generate the standard RGB script with generate_true_color_evalscript.
If you need to map vegetation health, run generate_ndvi_evalscript, which creates a ready-to-use script that colors output from dark tones up through deep green for dense plant life. For a false color view—which is common when mapping urban areas or water bodies—you generate the specific script using generate_false_color_evalscript, emphasizing vegetation in red, urban structures in cyan/grey, and water in dark blue.
Running the Job and Getting Results
When you've built your custom logic (your evalscript), it’s time to run the job. You execute the full analysis by calling process_image, specifying the script, the geographic area, and the date range. For statistical reporting over time, you use get_statistics to calculate summaries—like the mean, minimum, or maximum value of a specific band across an area of interest over defined periods.
This whole process lets you analyze data fields like NDVI and NBR on demand without ever downloading terabytes of raw files.
019dd157-b35f-7031-829e-f02624272bc8 
Here's how it actually works
The bottom line is: you define the area, generate the processing instructions, and run them all in one sequence to get an analyzed output.
First, use search_cloud_free or catalog_search to narrow down the dataset. You define the area and date range, ensuring you only work with high-quality imagery.
Next, if you need a specific visualization (like NDVI), run generate_ndvi_evalscript. This creates the necessary processing instructions (the evalscript) that tells the server exactly how to combine and colorize the bands.
Finally, execute the script using process_image, providing the generated evalscript along with the coordinates and date range. The result is a processed image or statistical report.
Who is this actually for?
Environmental scientists who need to track ecosystem changes; urban planners monitoring land use shifts; agricultural advisors needing real-time crop health reports. This is for anyone whose job depends on analyzing physical, measurable change across vast geographic areas.
Uses catalog_search and get_catalog_item to pull metadata on specific datasets; runs process_image to render custom band combinations.
Runs the full cycle: uses list_band_combinations to select an index, generates the script using generate_ndvi_evalscript, and calculates trends with get_statistics for ecosystem monitoring.
Uses search_cloud_free to find clean imagery over crops; runs NDVI analysis to determine if a region needs intervention.
What Changes When You Connect
Track Change Over Time: Use get_statistics to calculate mean, min, or max metrics across an area over daily, weekly, or monthly periods. This lets you measure gradual change—like deforestation rates—which is impossible with single-shot data.
Filter Out Bad Data: Don't waste time processing clouds. Run search_cloud_free first to guarantee your input imagery meets a low cloud cover threshold (<10% for clean analysis).
Generate Specific Maps: Instead of raw data, use tools like generate_ndvi_evalscript to get scripts that automatically colorize the output based on vegetation density. The result is an immediate, usable map.
Compare Different Viewpoints: Use list_band_combinations to see if a simple True Color view works, or if you need a False Color composite for better distinction between water and land.
Handle Massive Data Volumes: You never download terabytes of raw data. Everything is processed by the API in place, returning only the final, analyzed output map or metric set.
See it in action
Assessing wildfire damage (Emergency Managers)
A manager needs to know how bad a fire was. They ask their agent: 'Find all Sentinel-2 data for the burn zone last month, focusing on NBR.' The agent uses search_cloud_free first, then runs list_band_combinations to confirm the Burn Severity index, and finally executes process_image to deliver a map showing precise damage levels.
Monitoring crop health (Agricultural Advisors)
A farmer needs to know if their field is struggling. They ask: 'Run an NDVI analysis for the last 30 days in this county.' The agent generates generate_ndvi_evalscript, finds clean imagery, and runs get_statistics on the resulting time series data, giving a quantifiable metric of decline or recovery.
Mapping land use change (Urban Planners)
An urban planner needs to track where buildings went up. They request: 'Compare this area from 2015 vs. 2023.' The agent uses catalog_search to pull both time slices, and then runs a statistical analysis (get_statistics) comparing the mean spectral reflectance between the two periods.
Finding specific data layers (GIS Professionals)
A GIS expert only wants radar imagery. They ask: 'List all available Sentinel-1 and Landsat collections.' The agent uses list_catalog_collections to identify the correct source, then uses get_catalog_collection to confirm its exact structure before proceeding.
The honest tradeoffs
Treating all data as equal
Running a complex analysis immediately after running catalog_search. This fails because the search only returns metadata; it doesn't provide the processed image required for analysis.
Always check cloud cover first. Use search_cloud_free to filter out unusable data before generating scripts or calling process_image. This saves computation time and keeps results clean.
Missing the index generation step
Calling process_image without a specific processing script. The server won't know which bands to combine, resulting in an error or a useless default image.
If you want NDVI or True Color, run the dedicated generator first: use generate_ndvi_evalscript or generate_true_color_evalscript. This provides the required input script for process_image.
Overlooking the data source
Trying to analyze a specific type of index (like moisture) when the dataset only contains optical imagery. You'll get an error because the needed spectral band doesn't exist.
Check list_catalog_collections first. This tells you if your target data—whether it’s radar (Sentinel-1) or optical (Sentinel-2)—is available for the index you need.
When It Fits, When It Doesn't
Use this server when the core problem is analyzing complex, multi-temporal physical measurements across a large area. You MUST use this if your goal involves calculating indices (NDVI, NBR) or tracking statistics over time (daily/weekly averages).
Don't use it if you just need to know where a dataset lives; for that, list_catalog_collections is enough. Also, don't use it if your analysis requires atmospheric correction beyond the built-in evalscripts—that needs external scripting.
If you only want simple metadata (like coordinates and dates), stick to catalog_search. If you need a full scientific report comparing metrics across time, this server is necessary.
Questions you might have
How do I find cloud-free images using search_cloud_free? +
You provide a bounding box, date range, and your required maximum cloud cover percentage (e.g., 10%). The tool filters the entire catalog to return only scenes that meet that cleanliness requirement.
Can I run NDVI analysis without generating_ndvi_evalscript? +
No, you must generate the script first. generate_ndvi_evalscript creates the specific processing instructions needed for the process_image tool to know how to calculate and colorize the index correctly.
What is the difference between catalog_search and list_catalog_collections? +
list_catalog_collections gives you a master list of all available data sources (like 'Sentinel-2'). catalog_search lets you actually search within one or more specific collections by date and location.
How do I get statistics over multiple months using get_statistics? +
You must provide a valid evalscript that defines the index, then specify a time series aggregation (daily, weekly, monthly) along with your bounding box to calculate metrics like mean or standard deviation.
How do I check if my client_id:client_secret credentials are valid using the `check_sentinel_hub_status` tool? +
Running check_sentinel_hub_status confirms your authentication and service connectivity. This is the first step to ensure your AI agent can actually talk to the Sentinel Hub API without hitting credential errors.
If I want to know which bands are necessary for a specific index, should I use `list_band_combinations`? +
Yes, using list_band_combinations provides a comprehensive list of indices and specifies the exact bands required (e.g., B08, B12) and the target collection for each one.
When is it better to use `search_by_tile` versus general coordinates in a standard `catalog_search`? +
Use search_by_tile when you need precise, standardized spatial referencing typical of Sentinel-2 data (e.g., 33UUP). Bounding boxes are good for general areas, but tiles give you the specific MGRS grid alignment.
I found a potential item ID; how do I confirm all its available metadata using `get_catalog_item`? +
Running get_catalog_item pulls the full, detailed metadata for that specific STAC catalog entry. This confirms if the item exists and gives you data points like geometry, cloud cover, and band information.
What is an evalscript and how do I use one? +
An evalscript is a small JavaScript program that tells Sentinel Hub how to process satellite bands into an output image. It defines which bands to use, how to combine them, and what colors to assign. You can use the generate_ndvi_evalscript or generate_true_color_evalscript tools to get ready-made evalscripts, then pass them to the process_image tool.
Can I analyze vegetation health with this server? +
Absolutely. Generate an NDVI evalscript with the generate_ndvi_evalscript tool, then process imagery for your area of interest with the process_image tool. For time-series analysis, use the get_statistics tool with temporal aggregation to track vegetation changes over weeks or months. The search_cloud_free tool helps you find clean scenes without cloud contamination.
What is the difference between this server and the Copernicus Data Space server? +
The Copernicus Data Space server focuses on product catalogue search and download — finding and retrieving raw satellite data files. Sentinel Hub focuses on on-the-fly processing — rendering images, computing indices, and generating statistics without downloading raw data. They complement each other: use Copernicus for data discovery and bulk download, Sentinel Hub for real-time analysis and visualization.
We've already built the connector for Sentinel Hub. Just plug in your AI agents and start using Vinkius.
No hosting. No infrastructure. No complex setup.
All 14 tools are live and waiting.
You're up and running in seconds.
Gemini Vinkius gives your AI agents access to the full catalog of app connectors, all fully managed, secure, and enterprise-ready. One subscription, every tool you need.
Built, hosted, and secured by Vinkius. You just connect and go.