Step-by-Step: Uploading Your First Photo to Raven

So you've heard Raven can look at a photo and guess where in the world it was taken, and you want to actually try it rather than just take our word for it. Good — that's exactly what it's built for. The whole flow, from picking a photo to reading your result, takes about a minute, but it helps to know what to expect at each step, especially the parts that are easy to get wrong on a first try, like choosing a photo that doesn't give the AI much to work with.

Step 1: Pick a Good First Photo

The single biggest lever over how impressive your first result feels is the photo you choose. Raven's AI reads environmental context — architecture, vegetation, road markings, signage, light — so a photo with some visible outdoor scene gives it far more to work with than a tight indoor close-up or a photo of, say, a plate of food. Good candidates: an old travel photo you already know the real answer to, so you can check the guess against your own memory; a screenshot from a show or film with a distinctive outdoor setting; or a friend's photo from a trip they're being cagey about. Street scenes, coastlines, building facades, and hillsides all tend to produce genuinely interesting, checkable results.

Step 2: Sign In (and Why That Gate Exists)

Before your first upload, Raven will ask you to sign in with Google. This isn't about gatekeeping the rest of the site — you can read the blog, the FAQ, and everything else without an account — it's specifically there to keep the actual analysis feature usable and abuse-resistant, since each guess involves a real call to Google's Gemini model behind the scenes. Signing in takes one click through the standard Google OAuth flow and doesn't require filling out a profile or picking a password.

Step 3: Upload and Watch It Think

Once you're signed in, drop your photo onto the upload area or tap to browse for it. You'll see a brief scanning animation while the photo is sent off and Gemini's vision model works through it — architecture, plant life, light and shadow, any visible text or signage, road and vehicle details — typically finishing in just a few seconds. There's nothing to configure or tune here; the analysis is fully automatic once the image is uploaded.

Step 4: Reading Your Result

Your result comes back as a location guess — sometimes a specific city or region, sometimes a broader area if the photo is genuinely ambiguous — paired with a confidence indicator rather than a bare assertion. That confidence score is worth paying attention to: it's the AI's own honest read on how much the photo actually supports its guess, not a GPS coordinate or a claim of certainty. A photo full of distinctive clues — a recognizable roofline, a legible sign, a particular species of tree — earns a confident, narrow guess. A hazy landscape shot with a plain sky and generic vegetation will honestly come back vaguer, and that's the system working correctly rather than failing; a wide, low-confidence guess is a more useful answer than a falsely specific one.

What Happens to Your Photo Afterward

Nothing, on purpose. Raven processes your image only in memory long enough to produce that single guess, then discards it the moment the request finishes — it's never written to a disk, a bucket, or a database. That's true whether the photo is a cherished old memory or a random test shot, which makes it comfortable to try the tool on genuinely personal travel photos rather than only throwaway images.

If You Want This On the Go

Raven lives in a browser, which is great for revisiting old photos from your camera roll or laptop, but if the curiosity strikes mid-trip — looking at a scene through a train window and wondering exactly where you are — our sibling app Geospy AI brings the same idea to your iPhone, available on the App Store, so you don't have to wait until you're back at a computer to get a guess.

That's really the whole flow: pick a photo with something to look at, sign in once, upload, and read the guess alongside its confidence rather than as gospel. First-time visitors get a free guess to try before anything resembling a paywall shows up, so there's no real reason not to just go try it on that one photo you've always wondered about.