# INPE STAC API MCP

> INPE STAC API Satélites MCP gives you direct access to Brazil's National Institute for Space Research (INPE) satellite metadata. Query Earth observation catalogs—CBERS, Sentinel, Landsat—to find specific scenes, collections, and asset details using standard spatial and temporal filtering.

## Overview
- **Category:** government-public-data
- **Price:** Free
- **Tags:** satellite-imagery, stac-api, earth-observation, brazil-data-cube, remote-sensing

## Description

You need to know what’s up in the sky over a specific region of Brazil. This MCP connects your AI agent to INPE's massive data catalog. Instead of writing complex API calls or navigating dense government portals, you just ask natural questions: 'Show me all Sentinel-2 collections that covered São Paulo between January and March.' The system handles the technical heavy lifting, letting you discover which datasets exist, what they contain, and where to find the specific assets. It's built on the STAC specification, so it’s predictable and reliable. If your agent needs access to deep geospatial data like this, Vinkius makes it simple; just connect it once and start querying everything from a single dashboard.

## Tools

### list_collection_items
Lists specific individual assets within a collection using geographic or temporal filters.

### get_collection
Retrieves detailed metadata for a single, named dataset collection.

### list_collections
Outputs a complete list of all data collections available in the catalog.

### get_conformance
Checks and reports on the STAC catalog standards compliance.

### get_root_catalog
Fetches the main, overarching structure of all available data catalogs.

### search_items
Searches for assets across multiple collections using specific coordinates and dates.

## Prompt Examples

**Prompt:** 
```
List all available satellite data collections from INPE.
```

**Response:** 
```
I've retrieved the available collections from the INPE STAC server. You can access datasets such as 'CBERS4-WFI-16D-2', 'S2-16D-2', and 'LANDSAT-8-16D-2'. Which one would you like to explore further?
```

**Prompt:** 
```
Search for satellite items in the bounding box [-48, -16, -47, -15] for the year 2023.
```

**Response:** 
```
Searching across collections... I found several items matching your criteria. For example, in the 'S2-16D-2' collection, there are scenes from various dates in 2023 covering those coordinates. Would you like the specific asset links for one of these items?
```

**Prompt:** 
```
Get details for the collection 'CBERS4-WFI-16D-2'.
```

**Response:** 
```
The 'CBERS4-WFI-16D-2' collection is a 16-day surface reflectance composite from the CBERS-4 WFI sensor. It covers the period from 2014 to the present and includes bands like Blue, Green, Red, and NIR. Do you want to list the most recent items from this collection?
```

## Capabilities

### Catalog Discovery
Determine the top-level structure of the INPE catalog using get_root_catalog.

### Collection Listing
List every available data set or collection, like CBERS4 or Sentinel-2, with list_collections.

### Metadata Retrieval
Get detailed information about a specific data group using get_collection.

### Asset Pinpointing
Find exact satellite scenes that match specific geographic boxes and dates with list_collection_items.

### Cross-Search Querying
Run a search across multiple data collections to find any asset matching your criteria using search_items.

### Schema Validation
Verify the catalog structure against STAC standards with get_conformance.

## Use Cases

### Tracking deforestation changes
An analyst needs to compare land cover from 2018 vs. 2023 in the Amazon basin. They ask their agent to search for items across collections using search_items, specifying both years and a large bounding box. The agent returns multiple matching scenes ready for comparison.

### Developing a new visualization tool
A developer needs to know what data is available without writing code first. They use list_collections to map out all datasets (like CBERS4-WFI-16D-2) and then call get_collection for technical details on the top three candidates.

### Investigating a recent weather event
A researcher wants data from a specific storm that happened last month. They use list_collection_items, filtering by the exact coordinates of the event and the precise date range to pull only relevant assets.

### Verifying data source compliance
Before accepting a dataset for publication, the team lead uses get_conformance on the catalog. This confirms that all metadata structures adhere strictly to STAC protocol, saving time debugging broken schemas.

## Benefits

- Skip manual querying. Instead of writing a multi-step script to find an asset, you ask your agent to search across collections using search_items and get the full list in one go.
- Validate data integrity right away. Run get_conformance before committing to a workflow; it confirms that the catalog meets the expected STAC standards.
- Avoid guesswork when sourcing data. Use list_collections first to see every available dataset, then use get_collection for specifics on any single one.
- Focus only on what matters. Instead of browsing huge directories, you specify a bounding box and date range with list_collection_items to retrieve precise assets.
- Build complex workflows quickly. Your agent can chain calls: first using get_root_catalog, then drilling into specific collections via get_collection.

## How It Works

The bottom line is you tell your AI client what kind of satellite imagery you need, and it figures out exactly which collection holds that data.

1. First, subscribe to this MCP and supply your Brazil Data Cube (BDC) Access Key.
2. Next, prompt your agent to identify the available data collections using list_collections or get_root_catalog.
3. Finally, ask the agent to narrow down the results by providing a bounding box and date range using search_items.

## Frequently Asked Questions

**How do I find out what collections are available using list_collections?**
Simply prompt your agent to run list_collections. It will output a comprehensive list of all datasets, like CBERS4 and S2-16D-2, that you can then explore further.

**Can I search across multiple data sources using search_items?**
Yes, this is exactly what search_items does. You provide a bounding box and a time range, and it searches all relevant collections simultaneously for matching assets.

**What should I use to check the structure of the whole catalog? (get_root_catalog)**
Use get_root_catalog when you need the highest-level view. It provides the foundational map, showing all the major sections and how they relate before you drill down into specific data sets.

**Is there a way to validate if the catalog adheres to standards? (get_conformance)**
Yes, running get_conformance validates the entire catalog structure. It confirms that the metadata follows the necessary STAC specifications, giving you confidence in the data integrity.

**How do I narrow down my search for scenes using bounding boxes or dates with `list_collection_items`?**
You specify spatial and temporal filters in your request. This allows you to retrieve only the exact assets needed, rather than a full list of every item. It's essential for limiting results when monitoring specific geographical areas.

**When I run `get_collection`, what kind of detailed metadata can I expect about that data set?**
You get the deep technical details, including sensor specifications, band names (like Blue or NIR), and coverage period. This information helps you understand exactly what kind of scientific data you're working with.

**What should I do if my query fails when using `search_items` because the coordinates are outside the map area?**
The MCP will return a specific error message detailing why the search failed. You just need to adjust your bounding box parameters or check your coordinate system. The agent handles these failures gracefully.

**What kind of access credentials are required before running any query, like `get_root_catalog`?**
You must first provide a valid Brazil Data Cube (BDC) Access Key when setting up the connection. This key authenticates your request and allows the agent to start querying the catalog.

**How can I see what types of satellite data are available?**
Use the `list_collections` tool. It will return a list of all available datasets in the INPE Brazil Data Cube, such as CBERS-4, Sentinel-2, and various synthesized data cubes.

**Can I search for specific images using geographic coordinates?**
Yes! Use the `search_items` tool and provide a `bbox` (bounding box) string in the format '[minx, miny, maxx, maxy]'. This allows you to find all satellite scenes covering your area of interest.

**How do I get the technical metadata for a specific collection?**
Use the `get_collection` tool with the specific `collection_id` (e.g., 'CBERS4-WFI-16D-2'). It will provide detailed information about the spatial extent, temporal interval, and available assets for that collection.