This ChatGPT prompt can find where this photo was taken
AI Prompts
This Prompt Turns Any AI Chatbot Into a GeoGuessr Detective
Give an AI model a single photo and this prompt, and it will walk through the image like a seasoned GeoGuessr player: reading road signs, checking plate formats, comparing vegetation and architecture, and narrowing down a country, a city, and sometimes a street.
What this prompt actually does
Most people ask an AI “where was this photo taken?” and get a vague, overconfident guess back. This prompt fixes that. It forces the model to slow down, list every visible clue first, and only then commit to an answer, with a confidence score and a note on what would prove it wrong. That structure is what separates a lucky guess from an actual visual investigation.
The clues it’s told to look for
The prompt
Copy this exactly and paste it into any AI model that can read images (Claude, ChatGPT, Gemini all work), then attach your photo.
Act as an expert GeoGuessr player and visual geolocation investigator. I will upload a photo and your job is to determine where it was taken. Analyze the image carefully before making your conclusion. Look for: - Architecture and building styles - Street signs and written language - Road markings and traffic signs - Cars, license plates and driving side - Public transportation - Businesses and recognizable brands - Vegetation and climate - Terrain and geography - Street furniture - Utility poles and infrastructure - Landmarks - Shadows, lighting and sun position - Any tiny visual details that could indicate a specific region Do NOT immediately guess. First identify all geographic clues you can see. Then give me: 1. Most likely country 2. Most likely city/region 3. Most likely specific location or landmark 4. 3 alternative locations 5. Confidence level for each 6. The strongest evidence supporting your #1 guess 7. What evidence would prove your guess wrong If you recognize a specific landmark, business, street or building, say so. If the image does not contain enough evidence for an exact location, say that clearly rather than inventing details. Give me your final answer in this format: 📍 MOST LIKELY LOCATION: CITY, COUNTRY 🎯 CONFIDENCE: XX% 🔎 WHY: [brief explanation] 🗺️ ALTERNATIVES: 1. 2. 3.
How to use it
- Pick a photo with visible context. A close-up selfie won’t work. You need street, sky, signage, or landscape in the frame.
- Open a chat with an AI model that supports image uploads. Claude, ChatGPT and Gemini all handle this well.
- Paste the prompt first, then attach the photo in the same message so the model reads the instructions before it looks at the image.
- Read the clue list before the final answer. That’s where you can judge whether the reasoning actually holds up, rather than just trusting the city name at the bottom.
- Push back if something looks off. Ask the model to reconsider a specific clue, like a license plate shape or a road sign font, if you know that detail better than it does.
Why this is a good prompt to study
Beyond the GeoGuessr novelty, this is a solid template for getting more reliable answers out of any AI model on tasks that involve judgment. The pattern is simple: list the raw evidence first, rank alternatives instead of giving one answer, attach a confidence level, and state what would disprove the conclusion. You can reuse that same structure for image-based research, product comparisons, or fact-checking tasks that aren’t about geography at all.
Want more prompts like this every week?
Simply Hustler News of the Week rounds up practical AI prompts, tools and workflows, no hype, just what actually works.