Count a shelf with one photo.
gkalogs reads shelves, labels and voice commands, so clerks stop scanning every box. Built for gloves, forklift noise and Wi-Fi dead zones.
- Photo a shelf. Get the SKU and the count.
- Say the adjustment. Skip the keypad.
- Lose Wi-Fi. Keep counting.
SKU 459 · 12 counted
SKU 301 · 8 counted
SKU 872 · not expected here
Expected14
Counted12
Missing2
Bin A-14, SKU 459: 2 units short. 75% likely in adjacent Bin A-15.
Manual cycle counts burn labor hours.
- Clerks scan hundreds of barcodes every shift.
- Typing on a phone while holding a clipboard causes entry errors.
- Misplaced stock becomes phantom inventory: the system shows units the bin does not hold.
- Discrepancies reach management because clerks have nothing to trace them with.
Four tools built around the clerk's day.
Snap & Count
One photo of a shelf or pallet.
- Detects boxes and counts them.
- Reads labels and barcodes with OCR.
- Target: a 2-second photo replaces a 2-minute scan and count.
Voice logging
Push to talk. Hands stay on the load.
- Say: "Update Bin A-14: 12 units of SKU 459, two units damaged."
- gkalogs parses the bin, SKU, quantity and damage, then queues the adjustment.
- No typing. No looking down at a screen near moving equipment.
Misplacement alerts
Every mismatch comes with a place to look.
- Flags counts that differ from expected, and items that do not belong in a bin.
- Checks past movement data to rank where the stock likely went.
- Example: "75% likely in adjacent Bin A-15. Check the receiving dock."
Count routing
The shortest walk through your grid.
- Groups and orders the day's count tasks into one efficient path.
- Less walking. Less aisle congestion.
- Shifts finish earlier.
Built for the floor. Connects to your WMS.
On the floor
- Large, high-contrast buttons for gloved hands.
- Haptic buzz and an audio chime confirm every scan and voice command.
- Microphone tuned for forklifts and conveyors.
- Offline-first: photos, audio and counts stay on the device, then sync when Wi-Fi returns.
In your systems
- Runs in the browser on iOS, Android and rugged tablets. No app-store install.
- CSV export for quick hand-offs.
- REST API endpoints push counts to your existing WMS.
What we measure in a 90-day pilot.
| Metric | Manual today | gkalogs target |
|---|---|---|
| Time per bin count | 90 to 120 seconds | Under 15 seconds |
| Inventory accuracy | About 92% | 99.5% or higher |
| Data entry error rate | 3% to 5% | Under 0.5% |
| Discrepancy resolution | 15 to 20 minutes | Under 3 minutes |
Targets for a 90-day pilot. Manual figures are typical for cycle counting without AI.
Run a 90-day pilot in one warehouse.
Tell us about your site, your count process and your WMS.
Contact us