Mr.Count: AI Counter (Image)

The Onboarding Reality

Opening Mr.Count for the first time presents a clean, almost empty interface. No tutorial pops up. No walkthrough video plays. The app drops you directly onto a camera viewfinder with a single button that says "Capture." That is it. The 2500 free credits mentioned on the Play Store are not visible anywhere on the main screen. There is no counter showing how many credits remain. No settings gear icon is immediately obvious. The app assumes you already know what to do with an AI object counter. This is a stark contrast to most productivity apps that shower you with onboarding screens. Mr.Count trusts that its target audience — professionals in construction, retail, and logistics — already understand the workflow. Take a photo of a pile of lumber, and the app will count the pieces. Upload an image of a warehouse shelf, and it will tally the boxes. The interface does not hold your hand. It simply provides the camera and a capture button. That is the reality of the first ten seconds with this app. No frills. No distractions. Just a camera and a promise of AI-powered counting. The app does not ask for any permissions upfront. It does not request access to your photos or camera until you actually tap the capture button. That is when the permission dialog appears. This delayed request is a thoughtful touch. It does not bombard you with demands before you have even seen the app's purpose.

When Counting Meets Record-Keeping

The app offers two core features that seem straightforward on their own: high-accuracy object detection and the ability to save and export your counts. The Play Store description presents them as separate bullet points. But in practice, they interact in a way that is not obvious until you use them together. Take a photo of a mixed pile of construction materials — say, 200 bricks and 50 bags of cement. The AI detects and counts them. The number appears on screen. Then you tap "Save." The app records the count, but it does not ask you to label what you just counted. It does not prompt you to categorize the items as "bricks" or "cement." It simply saves a timestamped entry with a number. This becomes problematic when you export the data. The export function generates a plain list of numbers with dates. There is no context. If you counted bricks on Monday and cement on Tuesday, the exported file shows "150" and "75" with no way to tell which is which. The Save & Export feature and the High Accuracy detection model are designed to work together, but the missing labeling step creates a gap. The app accurately counts objects, but it does not help you organize those counts meaningfully. A professional managing inventory needs to know what was counted, not just the number. This is a real friction point that emerges only when you try to use both features in a real workflow. The app gives you the raw data but leaves the interpretation entirely up to you.

Performance Without the Noise

Mr.Count does not show ads. That is a notable detail in a free app. Many free productivity tools rely on ad revenue to sustain themselves, but this app appears to operate on a credit system instead. The 2500 free credits are the currency. Each count likely consumes a credit. The app does not display a credit counter, so you have no way of knowing how many counts you have left. This is a strange omission. If the business model depends on credits, the user should be able to track their balance. The app also does not crash during testing. It processes images in about two to three seconds, which is fast enough for on-site use. The computer vision model handles overlapping objects reasonably well. It does not get confused by shadows or slight variations in lighting. However, it struggles with highly cluttered scenes. A photo of a disorganized toolbox with dozens of small items produces inconsistent results. The app might count 45 items one time and 52 the next. This variability is not mentioned anywhere in the description. The app claims "high accuracy," but that claim is relative. For construction materials with clear edges and uniform shapes, the accuracy is impressive. For small, irregular objects in a messy environment, it is less reliable. The app does not offer a manual correction tool. If the AI miscounts, you cannot adjust the number. You have to retake the photo and hope for a better result. This lack of manual override is a significant limitation for users who depend on precise counts.

Permission Request and Data Handling

The app requests access to your camera and photo storage. This is expected for an image-based counter. What is less expected is the data safety section on the Play Store. The app collects personal information and photos and videos. It does not encrypt this data. It does not share data with third parties, but it also does not encrypt what it collects. The developer states that you can request data deletion. The app does not appear to use any trackers or analytics frameworks, based on the available AppBrain data. The permission request is straightforward: camera and storage. No location, no contacts, no microphone. The app's focus remains narrow. It does one thing — count objects from photos — and it asks for only the permissions required to do that. The Thailand-based developer has not published a privacy policy outside the Play Store. The data handling practices are described in the store listing, but there is no external website or document to provide more detail. This is a common pattern for new, small-scale apps. The permission model is clean, but the lack of encryption is a point of concern for users who handle sensitive inventory data.