CAPTD

Scene Recognition

Upload a photo and an AI vision model identifies the scene type, subject and lighting condition, plus what that combination is best suited for posting or shooting next.

This feature uses Artificial Intelligence to analyse your content and provide suggestions. Your images may be processed by an AI service depending on configuration. Please review our privacy policy.

  • Camera: the camera is used only for this photography tool and its AI analysis — it isn't activated automatically, and your browser will ask you to approve access the first time you use it.
  • Photo library: the photo you select will only be used for this analysis. Because it leaves your device for AI processing, we're telling you that now, before you choose one.
  • Storage: your photo is uploaded securely to generate this result, then deleted from our storage once the analysis completes. It is not kept in a history and is not used to train any AI model.
AI analyses this month0 / 30

30 analysises left — shared across Composition Analysis, Scene Recognition and Editing Suggestions. Resets next month.

How to use this

What it does

Sends your photo to Google Gemini's vision model to identify what kind of scene it is (portrait, landscape, street, wildlife, and so on), what the main subject is, the lighting condition, and what the shot would be best used for.

What you'll need

  • A photo (upload or take one with your camera).
  • Explicit consent, given via the checkbox above, since your photo is sent to an AI service for this one.

Steps

  1. Read and check the consent notice above.
  2. Upload a photo or take one with your camera.
  3. Tap "Analyze photo" and wait a few seconds for the result.

Reading the result

The scene-type badge is the AI's top-level classification; the confidence tag reflects how clear-cut the call was — a photo with an ambiguous or mixed subject may come back “Medium” or “Low” even with a reasonable guess. “Best suited for” is a suggestion for how to use the shot, not a rule.

Tips

  • This is a starting point for sorting or captioning a big batch of photos quickly, not a substitute for your own judgement on a shot you already know well.
  • This tool is capped at a shared 30 AI analyses per month across Composition Analysis, Scene Recognition and Editing Suggestions.

Learn more

What does scene recognition actually identify?

This tool sends your photo to Google Gemini's vision model, which classifies what kind of scene it is — portrait, landscape, street, wildlife, and so on — identifies the main subject, reads the lighting condition, and suggests what the shot would be best used for. It's a description of what the AI sees in the image, not a technical pixel measurement like the browser-based Photo Analysis tools.

Why photographers use it

Sorting or captioning a large batch of photos one by one, by eye, is slow — this does the first pass for you, useful after a big shoot when you're trying to quickly separate portraits from landscapes from candids, or when you just want a second opinion on a shot's strongest use case before deciding where to post it.

Common mistakes

  • Treating the confidence tag as an image-quality score. "Low" confidence means the scene itself was ambiguous or blended visually — a portrait shot in a dramatic landscape setting could plausibly read as either — not that the photo is poorly shot or badly focused.
  • Using this as a substitute for knowing your own shot. it's a starting point for quickly sorting a big batch, not a replacement for your own judgement on a photo you already know well and have a specific intent for.
  • Expecting fine-grained subcategories. it identifies broad, common scene types, not narrow specialisations — it may call something "wildlife" without distinguishing bird photography from big-game photography. Treat it as a first-pass sort, not exhaustive tagging.
  • Spending an analysis on a photo you already know the category of. this tool shares its 30-per-month cap with Composition Analysis and Editing Suggestions — save it for genuinely ambiguous shots or large-batch sorting, not photos whose category is already obvious.
  • Treating "best suited for" as a rigid posting rule. it's a suggestion based purely on visual content, not a platform-fit algorithm that accounts for your specific audience, personal style, or the caption and context you'd add around it.

Example settings

Tight, shallow-depth-of-field shot of a face

Close crop, blurred background, one clear human subject

Typically identified as "Portrait," high confidence, best suited for close, personal-connection content

Wide shot of mountains and sky, no human subject

Expansive scene, natural terrain, no people

Typically identified as "Landscape," high confidence, best suited for scenic/travel content

Candid mid-motion shot on a busy sidewalk

Unposed subject, urban setting, movement in frame

Typically identified as "Street" or "Event," sometimes Medium confidence if the scene reads as visually mixed

FAQ

How is this different from the free Photo Analysis tools?
Sharpness, Exposure and Colour Analysis measure technical pixel-level properties entirely in your browser, with no AI involved. This tool asks an AI vision model to describe and classify what the photo actually shows — a genuinely different kind of analysis, which is why it needs an actual model call and carries a usage cap and consent notice.
Is my photo stored after this analysis?
No — it's deleted immediately after analysis, not retained. See the Privacy Policy for full details on data handling.
Why did it give "Medium" or "Low" confidence for a shot I think is obviously one type?
Scenes that blend visual cues from more than one category — a portrait with a dramatic landscape backdrop, for instance — can genuinely read as ambiguous to a classifier, even when you have a clear intent in mind as the photographer.
Can I analyse a whole folder of photos at once?
No — one photo at a time, the same as the other AI Photography Assistant tools. There's no batch mode for this tool.
Does this tell me if my photo is good, or just what it is?
Just what it is, plus a rough best-use suggestion — it doesn't give a quality judgement on the shot itself. For that kind of feedback, try Composition Analysis or Editing Suggestions instead.