GeoInfer — AI Image Geolocation

🌍 Geolocate any image using AI — no GPS, no EXIF, no metadata required.

GeoInfer analyzes visual cues (architecture, terrain, vegetation, signage) to predict where a photo was taken, down to city level. Built for OSINT, digital forensics, investigative journalism, and security workflows.

Requires the GEOINFER_API_KEY environment variable.

Setup

1. 1. Get your API key from app.geoinfer.com/en/api
2. Run: INLINECODE1

Tools

Predict Image Location

bash scripts/predict.sh /path/to/image.jpg [model_id] [top_n]

• - model_id — model to use (default: global_v0_1). Run models.sh to list available models.
• INLINECODE5 — number of top predictions to return, 1–15 (default: 5)
• Accepts JPEG, PNG, WebP, and other common image formats (max 10MB)

Example:
CODEBLOCK1

Output: JSON array of top location predictions with coordinates and confidence scores.

List Available Models

bash scripts/models.sh

Check Credits

bash scripts/credits.sh

Credit Costs

Agent Usage Notes

• - Always run models.sh first if unsure which model to use
• Prefer global_v0_1 for general-purpose geolocation
• Use top_n=1 for fastest single-answer responses; top_n=5 for investigations requiring confidence comparison
• Check credits before running large batch jobs
• Images must be local files — download remote images before passing to INLINECODE11