Hey community!
I wanted to share a project I've been building and get some feedback from other developers working with spatial data or AI integrations.
The Problem & Idea
Extracting semantic insights from satellite imagery usually requires training custom computer vision models or setting up heavy GIS processing pipelines. For quick prototyping or lightweight apps, that felt like overkill.
I decided to build a lightweight REST API backend (hosted on Google Cloud Run) that pipes satellite/aerial imagery directly into Gemini's vision capabilities and returns structured JSON output.
How It Works
- Input: Accepts imagery or coordinates.
- AI Processing: Uses Gemini vision to analyze spatial features (land cover, change detection, object counting, etc.).
- Output: Formats the analysis into clean, predictable JSON ready for app integration.
Live Demo / Testing
I published the API on RapidAPI with a free tier so anyone can test requests directly in the browser playground:
Try the Satellite Vision API on RapidAPI
Looking for Feedback!
Since I'm actively improving the endpoints, I'd love to hear your thoughts:
- What imagery analysis tasks do you usually struggle to automate?
- Are there specific JSON parameters or formats you'd find useful for GIS workflows?
Drop your thoughts or edge cases in the comments below!
Top comments (1)
Here is the link to test it on RapidAPI: rapidapi.com/janikklimes10/api/sat...