Count fruit from a photo

Any fruit you can photograph: apples, oranges, mangoes, tomatoes, coconuts, melons and more. The engine is class-agnostic — it counts whatever repeated item fills the photo, and you can mark examples to pin it to one fruit.

Mark 1–3 example fruits by dragging a box around them and the AI counts exactly that item, ignoring everything else in the frame. Every counted item gets a dot on the image so you can verify the total at a glance.

It counts what's clearly visible, so open canopies with well-lit fruit work as a sampling tool for yield estimates. Heavily hidden fruit won't be seen — for precise totals, count at harvest in crates or trays.

Yes — counting photos is free with no signup for everyday use. Photos are uploaded only to process and are deleted automatically afterwards; we never keep or share them.

Similar items is built for a crate or bin of one fruit — all apples, all oranges — and gives the tightest count on a uniform harvest photo. Distinct objects switches to object detection so a mixed fruit bowl or market display can have its different fruits recognised separately. Auto is the safe default that picks between the two based on the photo.

Sensitivity (Low to High) sets how sure the AI must be before marking a fruit. Glossy skins like apples or citrus can create bright highlights that read as edges, so lower Sensitivity toward Low if it's over-counting from glare; raise it toward High if fruit at the back of a crate or in shadow is being missed. Diffuse, even lighting over the crate reduces how much adjustment you need.

Result overlay chooses how each counted fruit is marked — Dots + heatmap for both a per-fruit dot and a density map, Dots only for a clean annotated photo, Heatmap only to spot dense clusters. Ignore small detections raises the size cut-off so leaves, stems or small blemishes on the crate aren't mistaken for extra fruit — push it toward Large only if you're seeing an inflated count from debris.

Yes — switch Single photo to Bulk audit and upload a photo of every crate, tray or conveyor section. Each is counted with your Sensitivity setting, a grand total is calculated, and every crate's count is added to the tally sheet automatically, which suits harvest days with dozens of containers.

Add each crate's count to the tally sheet as you go, labelling rows like "Crate 12 — Gala apples", then export the whole run as a CSV for your yield records or a branded PDF report. Sign in to save the stock-take to your account and share a read-only link with the packhouse team.

Set Counting mode to Distinct objects so different fruit types in the same shot are told apart, keep Sensitivity moderate to avoid over- or under-counting on glossy skins, and set Ignore small detections low enough to keep small fruit like grapes but high enough to drop stems and leaves. Mark examples of one fruit type at a time and log each as its own tally row for a per-variety breakdown.

No — toggling from Single photo to Bulk audit so you can queue photos of every crate, tray or conveyor section at once is a Premium feature; the free plan counts one photo at a time, with each result added to the tally sheet manually. Counting mode, Sensitivity, marking examples and single-photo counting all remain free.

Yes — the tally sheet's CSV export is free for yield and harvest records, but generating the branded PDF stock-take report and using "Copy team link" to share a saved session with the packhouse team both require a Premium plan, prompting an upgrade if tried on the free plan. Saving a stock-take to your account stays free, though free accounts can only keep a limited number of saved sessions.

Tap Mark example items and drag a box around 1-3 examples of the fruit you want counted, then tap Done marking — each example can be removed individually with its own Remove example button, or Clear examples resets all of them at once. This is the fastest way to isolate one fruit variety in a photo that also shows crates, leaves or other produce, since the counter locks onto the marked example's appearance.