# EOSDA Agriculture MCP MCP > EOSDA Agriculture Satellite Data provides your agent instant access to global, high-resolution satellite imagery from sources like Sentinel and Landsat. It lets you calculate critical vegetation indices (like NDVI) or monitor soil moisture trends for any field worldwide. Instead of manually downloading massive data files, your AI client runs the whole process—from finding the right picture to providing a health score—in natural conversation. ## Overview - **Category:** the-unthinkable - **Price:** Free - **Tags:** satellite-imagery, precision-agriculture, ndvi, remote-sensing, crop-health, soil-moisture ## Description Need to check crop health across hundreds of acres? This MCP connects your agent directly to global remote sensing data. You can search for imagery from multiple satellites (Sentinel-2 and Landsat 8, for instance) covering specific date ranges or custom geographical areas. Once you have the right picture, the agent doesn't just show it; it runs complex calculations, like determining vegetation indices or monitoring soil moisture over time. Whether you’re optimizing fertilizer application or simply tracking land use change, your AI client acts as a dedicated remote sensing specialist through natural conversation. You connect this capability via Vinkius and keep all your geospatial intelligence in one place. It takes the guesswork out of fieldwork by providing measurable data points for every crop cycle. ## Tools ### search_dataset You search for satellite images covering a specific date range and location within a single, specified dataset. ### search_multi_dataset It pulls imagery from several different satellite missions (like Sentinel-2 and Landsat 8) simultaneously across the requested area and time period. ### create_vegetation_task This tool starts a task to calculate specific vegetation health metrics like NDVI or EVI for a given area of interest. ### get_available_datasets It lists all the satellite data sources available for searching imagery and running calculations. ### get_available_indices The agent provides a list of all possible vegetation indices you can run, such as NDVI or EVI. ### get_task_result This tool checks the status and retrieves the final processed data set from a previously initiated calculation task. ## Prompt Examples **Prompt:** ``` Find Sentinel-2 images for my farm from the last month. ``` **Response:** ``` Searching Sentinel-2 dataset for the specified date range... I found 3 cloud-free scenes. Shall I provide the scene IDs and download links for them? ``` **Prompt:** ``` Calculate the NDVI for this area: [GeoJSON coords]. ``` **Response:** ``` Initiating NDVI vegetation task for your area of interest... The task has been created with ID `task_123`. I will monitor the status and let you know once the result is ready. ``` **Prompt:** ``` What is the resolution of Landsat 8 satellite data? ``` **Response:** ``` Landsat 8 provides high-quality imagery with a spatial resolution of 30 meters for most bands. It has a revisit time of approximately 16 days. ``` ## Capabilities ### Identify available satellite datasets The agent retrieves a list of active satellite sources, including their technical specifications like resolution and revisit frequency. ### Search imagery across multiple satellites You instruct the system to pull scene IDs from several different satellite missions within a specified date range and geographical boundary. ### Calculate vegetation indices The agent initiates a processing task to calculate specific metrics, such as NDVI (vegetation health) or EVI (biomass), for your area of interest. ### Find imagery for one dataset You narrow the search down to a single satellite source and find all available images within a date range and specific location. ### Retrieve calculated results The agent pulls the final, processed data set from a completed vegetation index task, including download links and status. ## Use Cases ### Auditing regional soil moisture levels A researcher needs to compare dry-season versus wet-season soil moisture across three different regions. Instead of running three separate searches, they ask the agent to use search_multi_dataset and then run a vegetation index calculation task for all areas simultaneously. ### Investigating sudden crop stress A farm manager spots an unexplained patch of yellowing crops. The agent uses search_dataset to pull the newest available images for that exact GeoJSON area, allowing the manager to immediately assess if the issue started last week or months ago. ### Comparing multiple crop cycles An agronomist needs to compare yield estimates from a cornfield (using Landsat 8) versus a wheat field (using Sentinel-2). They use search_multi_dataset and then create separate vegetation index tasks for both types of crops. ### Checking dataset reliability A developer needs to know the best source for high-resolution data. The agent first calls get_available_datasets, which immediately provides technical specs like resolution and revisit times so they can build their application correctly. ## Benefits - Stop guessing about field problems. By running a vegetation index calculation task, you get precise, quantified data on plant biomass and stress levels for immediate action. - Don't search one dataset at a time. Use the multi-dataset search to pull imagery from multiple satellite missions simultaneously, giving you maximum coverage options in one query. - Know exactly what data is available before you start coding. The get_available_datasets tool lets you see every source (Sentinel, Landsat) and its technical specs upfront. - Track changes over time easily. You can search for imagery across a full date range, allowing you to monitor soil moisture trends or detect seasonal shifts in vegetation health. - Get the final answer without manual steps. After initiating a task using create_vegetation_task, the get_task_result tool handles waiting and downloading the completed data set. ## How It Works The bottom line is that your agent manages the entire process—from discovery to calculation to download—without needing you to manually call multiple API endpoints. 1. First, subscribe to this MCP using your EOSDA API Key in your Vinkius client. 2. Next, prompt your agent with the required location (GeoJSON) and desired analysis type. The system handles searching for available datasets and indices automatically. 3. Finally, once the index task is running, you use the dedicated tool to retrieve the processed result data set. ## Frequently Asked Questions **How do I find images from multiple satellites using search_multi_dataset?** You specify the name of several desired datasets (like Sentinel-2 and Landsat 8) along with your date range. The agent then collects scenes from all those sources into one result set for you. **What is the difference between search_dataset and search_multi_dataset?** Search_dataset looks only within a single, specific data source (e.g., just Landsat 8). Search_multi_dataset combines imagery from several different sources into one search result. **Does create_vegetation_task calculate soil moisture?** While the primary focus is on vegetation indices like NDVI, the system can perform tasks that monitor key environmental metrics including general soil moisture trends for your area of interest. **What do I use if my task fails? How does get_task_result help?** If a calculation task fails or is still running, you call get_task_result. This tool checks the current status and tells you whether the data is ready to download or if an error occurred. **Before I run any calculation, what credentials do I need for `create_vegetation_task`?** You must provide an API Key obtained from the EOS Data Analytics dashboard. This key authenticates your connection to the MCP and allows all tools, including dataset searches, to function correctly. **When using `create_vegetation_task`, what format is required for the area of interest?** The tool requires a GeoJSON object. This standard format lets you pinpoint the exact geographic boundaries for analysis, ensuring the calculated index only covers your specific farm or region. **Which indices are supported? How do I check available types using `get_available_indices`?** The `get_available_indices` tool lists every calculation type you can run. It specifies all metrics, such as NDVI and EVI, so you know precisely which ones to request for your crop health analysis. **What information does `search_dataset` give me about the satellite imagery I find?** The search returns critical details for each scene, including unique scene IDs, the precise date captured, cloud cover percentage, and direct download URLs. This helps you filter images based on quality or timing. **What satellites are covered by this integration?** The server provides access to Sentinel-2 (high resolution), Sentinel-1 (radar), Landsat 8 and 9 (historical and medium res), and MODIS (high temporal resolution). **How do I calculate the NDVI for a specific field?** Use the `create_vegetation_task` tool. You need to provide the `index_type` as 'NDVI', the `dataset_id` (e.g., 'sentinel2'), and the `aoi` (Area of Interest) coordinates in GeoJSON format. **Is the area of interest (AOI) required for searches?** For general searches, it is optional but highly recommended to narrow down results. For index calculation tasks (`create_vegetation_task`), the AOI is mandatory to define the target area.