OSINT techniques for beginners: how to investigate photos step by step

Open-source intelligence — OSINT — sounds like something reserved for government agencies and investigative journalists. In practice, it is simply the disciplined use of publicly available information to answer a question. And one of the most common OSINT questions is deceptively simple: where was this photo taken?

Whether you are a student exploring digital investigations for the first time, a small-business owner verifying a supplier's claims, or a hobbyist who enjoys the detective work behind GeoGuessr, this guide walks you through a beginner-friendly workflow for tracing photo locations. Every step uses free or low-cost tools, respects privacy laws, and produces findings that others can reproduce.

Step 1 — Preserve the source material

The single most important habit in OSINT is preservation. Before you analyse anything, download the highest-quality version of the image available. If it lives on a website, use the direct image URL rather than a screenshot. If it was sent via a messaging app, ask for the original file — messaging platforms often compress images and strip metadata during transmission.

Rename a working copy and keep the original untouched in a separate folder. This matters more than it sounds: if your findings are ever challenged — in a newsroom, a courtroom, or even a Twitter thread — you need to demonstrate an unbroken chain of custody from the original source to your analysis.

Step 2 — Extract and read metadata

Every digital photograph carries invisible data baked into its file structure. The most useful format is EXIF (Exchangeable Image File Format), which can include GPS coordinates, timestamps, camera model, lens settings, and even the software used for editing.

If the photo has intact GPS data, your job is essentially done — plot the coordinates on a map and verify the location visually.

The reality, though, is that most images shared online have had their EXIF stripped. Social media platforms like Instagram, Facebook, and Twitter remove metadata on upload for privacy reasons. Messaging apps do the same. So while EXIF is the fastest path to a location, you should expect it to be missing and have a backup plan ready.

Tools for extraction include ExifTool (command line, extremely thorough), Metapho (mobile-friendly), and PhotoRadar's built-in EXIF viewer. Even when GPS is absent, other fields — like timezone offsets, camera serial numbers, or editing software tags — can provide useful leads.

Step 3 — Scan the image for visual clues

When metadata is missing, the photograph itself becomes your primary data source. Train yourself to scan systematically rather than jumping to the most obvious feature:

• Text and language: Street signs, shop names, advertising billboards, and vehicle registration plates. Even a single word in a specific script — Cyrillic, Arabic, Hangul — eliminates most of the world.
• Weather and light: The angle of shadows indicates latitude and time of day. Overcast skies versus harsh tropical sun, snow versus dry pavement — weather is a surprisingly precise filter.
• Infrastructure: Road markings, traffic light designs, power line configurations, and socket types all vary by country.
• Vegetation: Palm trees, coniferous forests, terraced rice paddies, and desert scrub are all climate-specific. Combined with terrain features, vegetation can narrow the search to a specific biome.
• Landmarks and skylines: Distinctive buildings, bridges, monuments, and mountain silhouettes are often identifiable through map tools or local knowledge.

Step 4 — Use reverse-image search

Reverse-image search engines compare your photo against billions of indexed images to find visual matches. The results often include earlier uploads of the same scene with more context — a location tag, a caption with a place name, or additional angles that confirm the geography.

Run your image through all three major engines simultaneously:

• Google Lens — best general coverage.
• TinEye — best for finding the earliest known version.
• Bing Visual Search — often surfaces results the others miss.

If the image has been cropped or filtered, try searching with different sections — a distinctive building in the background may return results even when the foreground subject does not.

Step 5 — Pivot to maps

With a likely country, city, or neighbourhood in mind, switch to mapping tools. Google Maps and Apple Maps provide satellite and street-level imagery. OpenStreetMap offers detailed infrastructure data. Mapillary and KartaView provide crowdsourced street-level photos that cover areas Google has not reached.

The technique is straightforward: identify a distinctive feature in the photo — an intersection, a building façade, a bridge — and look for it on the map. Street View is particularly powerful for confirming matches, because you can virtually stand at the same spot and compare the view angle by angle.

Step 6 — Accelerate with AI

AI-assisted tools like PhotoRadar analyse skylines, terrain gradients, road geometry, and architectural patterns to propose candidate locations.

They are not a replacement for the steps above — they are an accelerant. Think of AI as a colleague who has memorised millions of geolocated reference images and can instantly suggest "this looks like coastal Portugal or possibly southern California."

The key is to treat AI outputs as leads, not answers. Verify every suggestion against maps and visual evidence before acting on it. Used correctly, AI can cut your investigation time from hours to minutes without sacrificing accuracy.

Step 7 — Document everything

The value of OSINT depends entirely on reproducibility. If someone else cannot follow your steps and reach the same conclusion, your findings are opinions rather than evidence.

Over time, this log becomes both a training resource and a defensible audit trail.

Legal and ethical boundaries

OSINT uses only publicly available information, but "publicly available" does not mean "anything goes." Data protection laws — GDPR in the UK/EU, state-level privacy laws in the US — apply to open-source research just as they do to any other data processing.

Stick to material that is genuinely public. Do not attempt to break paywalls, access private accounts, or contact individuals without a legitimate reason. The goal of OSINT is to make information clearer and more reliable — not to invade anyone's privacy.

Starter toolset

Every expert was once a beginner. The techniques described here are the same ones used by professional investigators and journalists — the difference is practice, not secret tools. Start with simple images, build your observation skills, and document your process.

For more structured playbooks, explore the investigator landing page or the journalist resource hub.