Skip to main content

Future of waste recycling and sorting with Lixo

At a time when European countries are increasing their recycling targets, particularly France with an ambition of 100% recycled plastic in 2025, the management and sorting of waste has now become essential more than ever. However, recycled material is often considered bad by industries due to its quality which is often substandard.


With this assumption in mind, Olivier Large and Marjorie Darcet created Lixo in 2019. This solution integrates a hardware and a software that helps waste handlers to analyse waste streams in real-time continuously. The objective ? Identify the composition of the stream for adapting the process of sorting and certifying the purity and quality of the sorted waste for better recycling.


Genesis of Lixo : KrabyTech

It all started in Paris, near the Saint Martin canal. Olivier Large, co-founder of Lixo, is appalled by the number of cigarette stubs in the canal. He, therefore, decides to create a small robot which looks like a spider and names it KrabyTech. Kraby sucks the stubs in the city by using a small module that identifies the stubs with images. Marjorie Darcet, co-founder, joins him in 2019.

I found his project interesting and the robot amusing but I did not consider myself an entrepreneur. Marjorie explains.

Rapidly, they realise that Kraby is difficult to commercialise and does not solve the problem of waste management on a bigger scale. However, Olivier and Marjorie understood that it was possible to expand the system of detection by computer vision to other types of waste. On that note, they decided to create Lixo with the aim to introduce artificial intelligence in the waste management industry which is mostly manual.

Arrival at the X-UP incubator

In September 2019, the Lixo team joined the X-UP incubator of the Ecole Polytechnique for six months : It was a real opportunity for us, the project was in its initial stages and the incubator provided us a framework especially since we first wished to go to the market and understand the problems of the sector before developing our product, says Marjorie Darcet

The waste management industry is very complex : it combines private and public, individual and collective all while being voluminous. The first four months of Lixo are hence dedicated to the exploration of waste management sector, meetings with collection agents, sorting centers, recyclers and municipalities for identifying needs. "Before developing anything, itw as important to check if our intuition was true about the facts and if there was some real economic and industrial interest for our product." indicates Marjorie

Olivier and Marjorie realise that the existing tools indeed do not allow proper follow-up of the quality and understanding of the waste flow. The industry has less quality control which is often only made on some ten kilos of waste manually. Why ? It is because of limited means. Waste is difficult to control because of its appearance ; a plastic bottle cannot be identified anymore once it has been crushed ! However, the recyclers wanted a permanent quality control for assuring the purity of the waste that they were buying to transform.

That’s when Lixo adds value with the identification of waste by computer vision. The waste management industry does not use lot of new technologies. It is a domain in which most actors invest less in R&D because they work with huge volumes of material at low cost. « A ton of paper only costs 30 to 40€ ! » adds Marjorie Darcet

How does Lixo work ?

Lixo is an artificial intelligence hardware and software solution that allows equips sorting machines, recycling chains and waste management systems. It’s machine learning algorithm treats continously in real time the flow of images for identifying waste composition. Its objective is to help managers to improve their resources and performance.


Lixo is a box that consists of a micro-computer, a camera and image transmission modules. The box captures images of waste flows and its image detection algorithm identifies the waste and its composition. To save electricity and data transfer, the algorithm is directly integrated in the micro-computer which avoids image transmission through Cloud. Only the result of waste identification is transmitted on the Lixo servers that retransmits this information on the dashboard of the clients. They can then know the flow rate (number of kilogram per hour), purity of the flow and distribution of the material for receiving sorting instructions and recalibrate the machines. The algorithm also identifies dangerous objects like inflammables to avoid fire.

Improve waste treatment

The principal benefit of Lixo is evidently the improvement of waste management for collectors and sorting centers. By continuous real time analysis, on a rolling carpet or a collection truck, it is easier to know the purity of waste and sort it. If we know that the quality is degrading, we can adapt the speed of the tapis or recalibrate the machine. Lixo adds a real value for quality control, with replacing the controller who remains the ultimate decision maker.

This is also a huge commercial benefit, adds Marjorie. The objective of the sorting center is to sell the material to a recycler. However, the recycler only buys as per a specific level of quality. With Lixo, the sorting center can attest the integrity of flow and improve the relation between the two actors.

Certification of nomenclature of waste

Another benefit of Lixo is that it serves recyclers who buy loads of waste for reuse. They need to be certain of the integrity of waste flows before buying from sorting centers. Lixo allows the certification of this nomenclature through artificial intelligence and continuous analysis of waste.

When we recycle PET plastic, the tolerance of indesirable material is between 1 to 10 PPM which means the tolerance limit is between 1 to 100 grammes in a ton of plastic bottles. It is extremely sensitive! At the same time, who would like to find a mix to glass and plastic in their recycled bottle? explains Marjorie Darcet

For the recyclers, a high rate of unwanted material results in loss of time, material but also money. For the flow of waste to function, it is necessary to be sure about the purity of material from the beginning. Hence, the advantage of using Lixo comes right from the point of collection.

Identification of sorting quality of collection trucks

The municipalities also have a role to play in waste management. By placing the Lixo boxes in collection trucks of recyclable waste, we can identify the sorting quality of citizens in real time. This allows the municipality heads to know the level of understanding of sorting instructions by people and sensitize them locality-wise with sorting ambassadors or adjust the frequency of collection rounds according to the quantity of waste collected.

The development of Lixo continues particularly through projects in Brittany and Centre-Val de Loire with the major players of the sector and municipalities. The team is also working with international prospects, especially in Switzerland. To carry out the projects well, the team of Lixo comprises of 6 people out of which four are data science engineers, mechatronics and web, for improving the performance continuously. To boost their development, Lixo est incubated at Agoranov and Wilco at Paris. Olivier and Marjorie shall launch their first fundraising programme in the coming months.