Fuli VisionFresh produce, seen.
Fuli Vision is built to see fresh produce the way a good grocer does: what it is, how fresh it is, and what is missing from the shelf.
Fuli Long builds computer vision for fresh-food retail in Southeast Asia.
Illustrative screen on a generated illustration. Confidence values and grades are examples, not measured results.
Five modules. One shop floor.
Our first product, Fuli Vision, starts at the weighing scale: a camera recognises loose fruit and vegetables so they can be weighed without barcodes or PLU stickers. Freshness grading, so stores can mark produce down before it spoils, is in development, and shelf-gap alerts from in-store cameras come next. Receiving checks and the owner's daily report follow on the same device.
Recognise at the scale
Loose produce named by the camera, confirmed with one tap. No PLU codes to memorise.
PrototypeRipeness and freshness
Graded on arrival and through the day, with a markdown time instead of a guess.
In developmentEmpty slots flagged
Shelf cameras tell the floor team what is missing, as it happens.
NextCheck every delivery
Weigh-in against the invoice, grade on arrival, photo evidence per line.
NextThe owner's evening report
What sold, what was marked down, what was thrown away, by item and supplier.
NextBuilt for the stores that weigh things.
Independent fresh-food stores first: many scales, thin margins, staff who change often. Then produce departments, online grocery picking, kitchens that receive produce every morning, and packhouses at the source.
Independent fresh-food stores
Every new cashier is productive on day one. Fewer mis-keys at the scale, markdowns on time.
The founder's two storesProduce departments and online picking
Self-service scales that recognise the item; one grading standard for every picker and every order.
NextKitchens and packhouses
Receiving checks where produce arrives by the crate; grading on the line before it leaves the highlands.
LaterAll five segments, why Malaysia first, and a waste calculator →
Built around two real stores.
We develop and test it in two Fuli Fresh Market stores in Kuala Lumpur and Puchong, co-owned by our founder, giving us real store images of local produce such as petai, kangkung and rambutan, varieties that global models rarely cover.
Fuli Vision is designed at the counter, not in a lab. It is built so staff confirm or correct each recognition with one tap, and every correction becomes training data for the next model.
Local produce has to be recognised where it is sold: open-front shops with mixed daylight and fluorescent light, items loose on the pan or bagged in plastic, bunches tied with raffia. That is the data global models do not have.
On NVIDIA RTX 4090 today. On Jetson in the store next.
Models in PyTorch. Store images stay in Malaysia, and recognition at the scale will not depend on a cloud round-trip.
Tested and running on an NVIDIA RTX 4090
We already test and run vision models on an NVIDIA RTX 4090 with CUDA at Universiti Malaya, through a research collaboration.
Training on local produce, inference in the store
Next: NVIDIA TAO Toolkit for transfer learning on local produce, then in-store inference on NVIDIA Jetson with TensorRT and DeepStream, so each store works without a cloud round-trip.
Status Recognition at the scale: Prototype · Freshness grading: In development · Shelf monitoring: Next · How it is built and measured →
Decided at the counter.
Fuli Long is a computer vision company founded by the co-owner of two fresh-food stores. The person who signs for the deliveries also decides what the product has to do. About the company →
Yang Li, Founder & Director
Runs two fresh-food stores in Kuala Lumpur and Puchong. Knows the produce, the staff and the scale, and started Fuli Long to build the tool the stores need.
FULI LONG SDN. BHD.
A private company limited by shares, incorporated in Kuala Lumpur under the Companies Act 2016. Fuli Vision is its first product.
- Legal name
- FULI LONG SDN. BHD. (富利龙)
- Registration no.
- 202601038796 (1700891-X)
- Incorporated
- 23 September 2026 · Companies Act 2016
- Headquarters
- Kuala Lumpur, Malaysia
- Product
- Fuli Vision · computer vision for fresh-food retail
- Stage
- Prototype in development · developed in two stores in Kuala Lumpur and Puchong
- Contact
- yangli@fulilongai.com
Straight answers.
What does Fuli Vision do?
Fuli Vision is computer vision for fresh-food stores. A camera at the weighing scale recognises loose fruit and vegetables without barcodes or PLU stickers (prototype). Freshness grading, so produce is marked down before it spoils, is in development; empty-shelf alerts come next. Built by Fuli Long Sdn. Bhd. in Kuala Lumpur.
Does it need barcodes or PLU stickers?
No. The camera recognises the item on the scale and proposes it; staff confirm with one tap. Items the model is unsure about fall back to the normal manual flow.
Where does it run?
Today we test and run vision models on an NVIDIA RTX 4090 with CUDA at Universiti Malaya, through a research collaboration. Next, each store gets a small NVIDIA Jetson device at the scale, so recognition works without a cloud round-trip.
Where do store images go?
Store images stay in Malaysia. The recognition demo on this site runs entirely in your browser; nothing you upload leaves your device.
Who is behind it?
FULI LONG SDN. BHD., a private company incorporated in Kuala Lumpur on 23 September 2026 (registration 202601038796). The founder, Yang Li, co-owns the two fresh-food stores where the product is developed.
Does it work with my existing scale and POS?
That is the design. Fuli Vision sits alongside them: the camera proposes the item and the staff confirm, and the existing label scale or POS prints the label and charges as before. Integrations will be built per store, starting with the scales in the founder's two stores.
Does it watch customers?
No. The cameras point at the scale pan and at the shelves, not at people. Faces are not what the models look for, and store images stay in Malaysia.
Can my store use it?
Not yet. Fuli Vision is a prototype in development. We are building it around the founder's two stores first; independent fresh-food stores in Kuala Lumpur and Selangor come next.
Run a fresh-food store in Kuala Lumpur or Selangor?
We are looking for the first stores beyond the founder's two: a scale with room for a camera, a normal week, and staff willing to tap confirm.