image

Image Counter

Manually count objects in photos with point or rectangle markers while preserving counted locations, review flags, and multi-image progress for export.

View on GitHub

Local processing

About your filesFiles and content selected in this tool are processed in your browser and are not uploaded to a Browser Kitty backend.

Operation demo See the basic workflow in a short video.

About this tool

Image Counter is a browser tool for counting objects in photos manually rather than asking AI to detect them automatically. It records not only the total but also where each item was counted using point or rectangle markers, making missed items and double-counting easier to spot in dense images.

Switch between multiple count categories, flag uncertain items for later review, and track completion across several images. Comments, capture-time information, and non-destructive display adjustments can be kept with the work, then exported as a read-only HTML viewer or as CSV plus marked JPEGs. Images and project data are not normally sent to a server.

Good for

Count parts, seeds, cells, products, equipment, or other visible items by human inspection
Keep a positional record of which objects were counted instead of storing only a final total
Count several object types in one image and return later only to uncertain markers
Pause a multi-image counting project and export the final results as CSV and marked images

What it can do

Key capabilities available in this tool.

Manually count by tapping point markers or drawing rectangles while keeping a visible record of counted locations
Move points, move/resize rectangles, change categories, delete markers, and use Undo/Redo
Flag uncertain objects for review and keep images incomplete while review items remain
Process up to 30 images with per-image completion, comments, and capture-time information
Non-destructive brightness, contrast, grayscale, and 90-degree rotation adjustments plus camera capture
Export a read-only self-contained HTML viewer or ZIP containing markers.csv, summary.csv, and marked JPEGs

How to use

Add JPG/JPEG, PNG, or WebP images, or capture a photo with the camera where supported.
Choose point, rectangle, or move mode and select the count category when you have more than one type.
Tap an object in point mode or drag around it in rectangle mode.
To correct a count, select an existing marker and move, resize, reclassify, or delete it. Flag uncertain markers for review.
Add comments or non-destructive display adjustments as needed, mark a finished image complete, and move to the next unfinished image.
Export a read-only HTML viewer or a ZIP containing markers.csv, summary.csv, and JPEGs with the markers drawn.

Supported

Input: JPG/JPEG, PNG, and WebP, plus in-browser camera capture where available.
Projects: up to 30 images with point, rectangle, and move modes, multiple count categories, and per-image completion/review/comments.
Display adjustments: brightness, contrast, grayscale, and 90-degree rotation without overwriting the original image.
Draft storage: automatic IndexedDB saving, with an option to WebP-compress only the draft image copy.
Output: self-contained read-only HTML viewer or ZIP with markers.csv, summary.csv, and marked JPEGs. A viewer HTML can be loaded back into the app to resume editing.

Limitations and notes

AI and automatic object detection are intentionally not used; the user decides what each object is and performs the count.
A project is currently limited to 30 images, each image to 35 MB, with an additional 220 MB guard for original image data across the project.
Large numbers of high-resolution images can consume significant device memory and browser storage.
Live camera capture depends on browser permissions and secure-context requirements; use the device camera or file input when unavailable.
EXIF capture-time reading primarily targets JPEG; other formats normally use file timestamps.
Version 1.0.0 does not export directly to Excel (.xlsx).

Frequently asked questions

Does Image Counter automatically detect and count objects?

No. It is a manual counting tool where you place points or draw rectangles. The priority is keeping a visible record of what a person counted rather than automatic detection.

Are images uploaded to a server?

No. Image loading, marker editing, camera processing, draft storage, display adjustments, and export run in the browser. The hosted version only requires the initial HTML download.

When should I use points versus rectangles?

Points are faster for simple counts. Rectangles are useful when objects are dense or small, or when you want to preserve the visible extent of each counted item for later review.

Can I stop and resume a counting project later?

Supported browsers auto-save project state in IndexedDB. You can also export a viewer HTML and load it back into Image Counter later to resume editing.

Offline version and source code

The standalone HTML has no runtime third-party dependencies and can be opened directly for local image, marker, draft, and export workflows. Camera features remain subject to browser permission rules.

View source on GitHub