# SatQuery AI

> SatQuery AI is a free web app that measures how a place on Earth changed between two dates, using Sentinel-2 satellite imagery. You place a square on a map and pick two dates; the app reads the imagery in your browser, measures vegetation, water and built-up land on both dates, reports the change in km², and explains the result in plain English. You can then ask follow-up questions. It was built for Smart India Hackathon problem SIH26167 (ISRO, Space Technology).

Key facts:
- Live app: https://satqueryai.cc.cd/ (free account, no API key, nothing to install).
- Imagery: Sentinel-2 L2A from Earth Search (C1) and Microsoft Planetary Computer; Landsat 8/9 is used only to fill cloud gaps.
- Analysis runs in the browser: a Web Worker streams Cloud-Optimized GeoTIFFs; there is no Python or raster backend.
- Output: seven spectral indices (NDVI, SAVI, NDWI, MNDWI, NDBI, BSI, NBR), five-class land cover, numbered change regions of at least 1 ha, and eight rendered views.
- A typical analysis takes about 25 to 55 seconds on a desktop browser.
- The AI explanation is written by Gemini 3.1 Flash Lite from the measured statistics and rendered views; it is told not to invent numbers.
- Limits: two-date comparison only, no exports, no offline mode, desktop browsers only (phones cannot hold two decoded dates in memory).

## Docs

- [About SatQuery AI](https://satqueryai.cc.cd/about/): what it does, how it works, the seven indices, data sources, privacy, limitations and FAQ.
- [Source code and README](https://github.com/p04pranav/SatQueryAI-SIH): architecture, the science behind each index, performance and reliability notes.

## Optional

- [Sitemap](https://satqueryai.cc.cd/sitemap.xml)