Corobox
Computer-vision quality control on the line
Hardware & IoTB2BAndijan
Corobox builds Edge-AI quality-control software for production and packaging lines. A cobot mounted on the conveyor uses computer vision to detect and remove defective products within seconds. Plants rent it as a service (RaaS), with no large capital outlay, and the line is automated in 10 days.
Traction
10
Factories approached, Sep
10 days
Integration time
$0
Upfront cost to the plant
- Admitted to the Founders Hub ecosystem
- Waitlist page live; pilot price and duration cut (Sep 2026)
The problem
Quality control at local plants still relies on people, and worker fatigue lets defects reach the market. Foreign automation is very expensive and halts the line for weeks, a risk directors won't take.
- Quality control relies on people, and fatigue lets defects through
- Foreign automation solutions are extremely expensive
- Installation halts the line for weeks
- Plant directors cannot take on that risk
The solution
A ready-made plug-and-play device for existing conveyors, not a bespoke project. No $20K+ upfront: plants pay a monthly fee per line, and integration takes 10 days instead of a month.
- Edge-AI - Runs on the conveyor, no cloud server
- Real time - Defects detected within seconds
- 10 days - Integration time, down from a month
- Plug-and-play - A ready-made device for existing lines
The company
What makes it different
Ready-made, not a bespoke project
No $20K+ upfront: plants pay a monthly fee per line under RaaS. The device ships as a ready box to any plant and mounts on an existing conveyor, cutting integration from a month to 10 days.
Who it's for
- Primary: food and packaging plants
- Testing ground: 1,500+ enterprises in Uzbekistan, to prove the product and show traction
How it makes money
- Hardware and software as a service (RaaS): monthly subscription per active conveyor line
- Additional: usage-based pricing
- No CapEx risk for the plant; high-margin recurring revenue for Corobox
Market
$10B+
Global visual inspection market
50,000+
Factories: Turkey, E. Europe, C. Asia
Cheaper Edge-AI chips let complex vision models run in real time on a compact device on the conveyor, with no cloud servers. Once proven locally, the box ships by ordinary logistics to factories in Turkey and Eastern Europe.
Competitors
- Manual labour - Human inspection at factories, the current de facto standard.
- Technical competitors - Integrators of costly foreign optical inspection (Cognex, Keyence) and local CV projects such as OnVision.
Founders
Hojiakbar writes all of Corobox's code as solo founder and full-stack engineer. Built MedAc, an AI assistant for doctors, from scratch through the innoX incubation programme, and Notedaily, a workspace platform, alone.
Studied Car Engineering at Seojeong University in South Korea, useful for a product that is a physical device on a factory conveyor. Also worked as a developer at the startup Offr and co-founded the Finatlas financial community.
The plan: bring in engineers who are stronger on the technical side, and put full focus on closing B2B contracts with factories and on operations.
Background
- Seojeong University, South Korea: Car Engineering
- Built MedAc, an AI assistant for doctors; graduated from the innoX incubation programme
- Built the Notedaily workspace platform alone as a full-stack engineer
- Developer at the startup Offr
- Co-founder of the Finatlas financial community
- Admitted to the Founders Hub ecosystem
Although I have a technical base, my main strength is project management and execution speed. I will not slow the company's growth by building the project alone; on the contrary, I bring engineers into the team who are far stronger technically than I am.
Milestones
- 4 SepSpoke with 10 factories; warm sales introductions through Otabek Suvonov
- 6 SepWaitlist landing page live; pilot price and duration cut; 4 cold calls; LinkedIn post aimed at MENA
- 10 SepComputer-vision build started; market research across 2 factories plus cafés and restaurants
- 22 SepFeatured on UzCombinator's official channel: defect spotted in a second, 10-day integration, $0 upfront
The round
$45,000
raising
Use of funds
Manufacturing the first batch of devices
45%
Product and Edge-AI model development
30%
Factory pilots and sales
25%
