LAY3RS’ mission is to safeguard natural and human heritage thanks to AI & blockchain.

Why should I care?​

Get involved and become an actor into the preservation of monuments, natural sites, flora, fauna, know-how and fictional environments.

How can I get involved ?

Join a digital twin project by contributing to its creation or evolution and become its curator. Decide on its developments and be rewarded for your contribution.

LAY3RS AI powered digital twins platforms

We are planning the development of several platforms dedicated to specific types of heritage and their specific user communities. Thus, it becomes possible to continuously expand the use cases of the LAY currency, to contribute to the preservation of an increasing variety of heritage subjects and to respond more specifically to technical specificities linked to intrinsic characteristics of subjects.


Monuments generated from heterogeneous databases to travel through time.


Most emblematic sports sites and events giving them a new posterity.



Places, Minerals, Flora and Fauna including the most beautiful natural sites and their components to procedurally generate evolving environments. 



Gestures and know-how to create digital beings capable of interacting and modifying their virtual environment.



Digital twins of fictional fragmented environments to generate global thematic environments.


What's next?

Would you like to find out more about Lay3rs?

Read all about it in our whitepaper.

LAY3RS supporting Partners

LAY3RS powering Technologies

Digital twin

Digital twin : A set of data relating to a subject to be preserved. It is therefore a dynamic set (evolving over time) of heterogeneous content (images, archives, reference 3D models, exports, acoustic map, semantic breakdown, etc.).

Export : A selection of data (e.g. a 3D model) optimized for a specific use (AR / VR / Unity / Unreal / Voxel, etc.). These exports can concern the entire subject or a specific part of it (work of art, architectural element, texture, etc.) and can be distributed on our platform and on various asset stores.

  • To equip researchers and architects in their preservation and renovation missions.
  • To visit a monument remotly or beyond the limits of time.
  • To serve as virtual studio sets or special effects for the creation of films, TV programs, advertisements or virtual events.
  • To anchor video games environment or metaverse in a specific context.
  • To create material reproductions (e.g. a 3D printing) of a subject or parts of it.
  • To support new forms of digital and immersive art.
  • To support all services located indoors or outdoors (with or without GPS signal).


And the best use cases have yet to be invented…

Enhanced archive

To keep a rich record of what is valuable and fragile by aggregating a large smart set of data on a specific heritage element.

Augmented mediation

To give the public a richer, more diversified, more personalized and more sustainable access to knowledge on a subject.

Economic model

To create a virtuous economic model specific to the twin, allowing to finance the preservation of the real heritage.

Support creation

To provide operational means for creators to reinterpret reality and history in new creative environments (VR, AR, Virtual Studios, metavers, etc.).

Emerging innovation

To propose a bottom-up innovation model to heritage institutions that creates a bridge between creators and them.

Massify the creation of digital twins

New environments of creation (XR) represent huge opportunities to invent new ways of promoting and broadcasting culture BUT production of a digital twin remains a very manual work process, very long, expensive, not very resilient (one capture for one output) and only based on visual perception.

In 2021, the European Union asked its member states to create 16 million digital twins of monuments by 2030. How to make this recommendation realistic without impoverishing it ? Lay3rs collaborates with research laboratories (CEA, ULB, Ecole Polytechnique) to develop new solutions to make this process more intelligent and able to fully exploit all existing data.

Heterogeneous data

Reconstruction of scenes and objects from heterogeneous data based on computer vision (SLAM) and artificial intelligence (NeRF) algorithms.

AI Generation

Production of 3D assets by artificial intelligence (reconstruction, generative and procerural).

Multimodal twin

Multimodal twinThe production of sound scenes based on a 3D sound engine and dynamic twins.

Since 2021, we have been working with the intelligent systems department of the CEA (French Atomic Energy Commission and main French applied research center) on various technologies for capturing and reconstructing 3D environments. Today, LAY3RS is considered as a startup « backed » by the institution. In fact, we are working daily with different laboratories which are giving LAY3RS access to a panel of expertise as specific as rare. This already productive partnership will continue to develop in the coming years, to allow LAY3RS to reach the most ambitious, innovative and valuable technological ambitions.

In addition to their use for the renovation or scientific study of a subject, when contributors invest in the creation of a twin, in the acquisition of new data, part of these funds are donated to the institution in charge of preserving the real subject. In time, the portion allocated to them by the DAO from the sale of licenses to exploit the exports will constitute a new channel of regular income for the latter.


Neural Radiance Fields, or NeRFs, are a recent deep-learning technique in visual computing that allows 3D scenes to be represented using neural networks.

At the heart of NeRF is the idea of representing a 3D scene not as a set of surfaces or voxel grid, but as a continuous field of radiance information. It’s a bit like imagining that the air around us is filled with tiny particles that have color and brightness properties, and these properties change depending on where and in which direction you look.

  • Deep Neural Network: NeRF uses a deep neural network to model this field. For a given point in space (defined by its x, y, z coordinates) and an observation direction, the network predicts the color and density which corresponds to transparency of the light passing through that point.
  • Volume Density: Instead of focusing on surfaces or voxel grid, NeRF views the entire space as a volume. Each point within this volume has an optical density.
  • Integration Along Rays: To generate an image, NeRF cast rays from the camera through the scene. For each ray, it samples several points along its path, uses the neural network to get the color and density of each point, and then combines these pieces of information to produce the final color of the pixel in the image.
  • Learning from Data: To train the network, we generally use a set of data (photos, videos, maps, etc.) of the scene taken from different angles. The network learns to reproduce this data and, once trained, can generate images of the scene from new angles and then transform these different views into a 3D model.

NeRF offers high-quality rendering, especially concerning the handling of fine details and complex lighting effects like refractions and reflections. However, they require substantial computational power and can be slow to generate. A lot of work is in progress in order to reduce the need of computational power and to accelerate the training and rendering.


Web3 models make it possible to invent new forms of public engagement and funding for common projects such as heritage preservation. Web3 tools are often designed for technological or financial uses, and notions of ReFi and digital commons are only just emerging.


How can we propose an ecosystem of decentralized tools that align interests and bring together audiences (web3 friendly or not), institutions and creators? 

A fungible token is a type of cryptographic token that is interchangeable with other tokens of the same type. This means that each unit of this token type is identical to every other unit. Fungible tokens are often used as currency because their interchangeable nature makes them useful for transactions.


LAY is a ERC-20 token for all services and platforms offered by LAY3RS. It is therefore used to :

  • Contribute to the creation of digital twins (finance data acquisition or produce 3D models).
  • Purchase licenses to use exports in all types of digital creations.
  • Automate the redistribution of products to complex and dynamic rights-holder structures.
  • Weight the powers of CAD contributors and give them access to the digital twin when they are involved in a project.
  • Participating in all the community mechanics and joining the marketing events.


While the token will be available to the public Q1 2024, a tesnet called PLAY (Prototype du LAY) is already available on the platform.


Total supply : 1 000 000 000 LAY

Technology : Polygon

A non-fungible token (NFT) is a type of cryptographic token that is unique and non-interchangeable with other tokens. Each NFT has information or attributes that make it different from other tokens, which can give it a different value.


By assigning to each set of raw data (ERC-721) and to each export (ERC-1155), it becomes possible to manage complex rights structures and to pay fairly (over the entire value chain) and transparently for the various revenues generated by the distribution of exports.


A Decentralized Autonomous Organization (DAO) is a form of organization that is managed by rules coded as smart contracts on the blockchain. A LAY3RS’ DAO offers to contributors the ability to vote and handle its finances through web3/blockchain technologies for creation or development of a digital twin.

A DAO is collectively owned and controlled by its members, rather than by a single entity or a small group of individuals. In a DAO, all decisions are made through consensus or voting, with each member having a proportional vote based on their ownership stake in the organization. The governance rules of the DAO are encoded on the blockchain, which means they are transparent and immutable.


DAOs can be used for a wide range of purposes, from managing digital assets to organizing collaborative workgroups. They are designed to be fully transparent, thereby avoiding issues of corruption and conflicts of interest that may arise in traditional organizations.


Initial partnership with CEA List
July 21
Winner of the Future Investment Plan - France 2030
September 22

Matrice (Social & Technological Institute) presents the project to create an ad-hoc structure to develop Lay3rs services.

Company creation
November 22
Build3rs Alpha Launching
July 23
Build3rs Public Launching
September 23
PLAY (Prototype of LAY) Tesnet Launching
September 23
DAO Features
October 23
Assets Marketplace features
November 23
NeRF-based 3D Reconstruction
January 24
Data Hunting Features
January 24
LAY Token Generation Event
January 24
Dynamic NFT
June 24
NeRF Editing Tools for 3D Artists
September 24
Account Abstraction
September 24
Digital twins library Desktop tool
October 24
Generative modelling of 3D
March 25
Human Reconstruction & Tracking
April 25
Universal NeRF
July 25
Multi and Hyper Spectral NeRF
September 25


Sébastien Malcotti


Yann Toullec


Gamora Yu


Xavier Aubert

CTO Digital Twin

Benjamin Jornet

CTO Blockchain



Sébastien Casaert

Head of Product

Rebecca Cervasio

AI Researcher 

Victoriya Kashtanova

AI Researcher

Zhong Qiang

Commercial manager

Mateusz Baranowski

Lead community manager

Mathieu Da Silva


Môsieur Pierre

Asset managment

François-Xavier Petit

Public Partnership Advisor

Emmanuel Ea

Development Advisor

Adrien Basdevant

Web3 Legal Advisor