AI creates greener and cycle-friendly streets


The challenge

The Netherlands is the cycling country par excellence. With an average of 1.3 bicycles per inhabitant and more than 37,000 km of cycle paths, cycling is an important part of Dutch culture and identity. For Dutch people, cycling is not just a hobby, it is how they live.

Cycling is good for your health and the environment. The Netherlands has therefore put the bicycle at the center of its plans for a climate-neutral future. The Netherlands Board of Tourism & Conventions (NBTC) has the ambition to inspire cities worldwide to follow the Netherlands' example and create more space for bicycle-friendly environments, to promote sustainable and healthy mobility.


To inspire the world, we have developed an AI tool to help people see in mere seconds what their cities could look like if they were designed for pedestrians and cyclists rather than cars.

With the help of artificial intelligence (AI), Google Streetview and the knowledge of Dutch cycling experts, people worldwide have the opportunity to add a touch of the Dutch cycling landscape to their own living environment.

Visit the AI tool

Want to work with us?

From busy street to bicycle-friendly in realtime

The Cycling Lifestyle AI tool we developed makes every street greener and more pleasant in an instant. For this purpose, various AI tools that had not previously been brought together were combined. The tool turns streets around the world into typical Dutch streets in real time, with lots of greenery and space for cyclists. The user is then shown four variations. Finally, the user is encouraged to share the street, showing others the miraculous make-over.

Code Sharing

The Dutch cycling lifestyle tool relies mainly on a technology called Stable Diffusion. This technology allowed us to generate images using text prompts on a model that's trained on a lot of images. It also allowed us to "paint in” a certain area of an existing picture. To determine this area we used a segmentation model that detects the road and vehicles in the original images.

We made the tool available for free to developers so that everyone in the world can help make streets bicycle-friendly.

We made the AI backend open source by merging it into a single ComfyUI workflow. ComfyUI is an open-source community framework for creating generative-AI workflows.  Next to all the models we used, we provided all the necessary documentation to enable any developer to recreate the AI-tool with clear inputs and outputs.

Check our code

split image

AI tool shows potential worldwide

The tool was a success: within a month it was used more than 1 million times. 
From New York to New Delhi: streets all over the world were given a bicycle-friendly make-over and then shared on social media. People shared their experiences and enthusiasm, demonstrating the great vizualisation power of the tool. 'How we design our streets is a choice. We can make different choices,’ Welsh Senate Member Lee Waters said on Twitter/X. Other users shared the results with their friends or local initiatives, with comments like: ‘Fabulous, I wish my street looked like this!’ and ‘Fantastic. I’ve already sent the results to my local councillor’.


Fullscreen image

A global inspiration call to cycle with Dutch flair

Complementing our digital efforts, a physical campaign strategically placed Out of Home (OOH) ads in key cities—Brussels, Dusseldorf, London, and Paris. Vibrant billboards, located on busy routes, featured a dynamic 'before' and 'after' transition on digital screens, sparking interest in our AI tool. A call to action directed viewers to, where a three-week pop-up served as a metric gauge for campaign effectiveness.

Fullscreen image
Fullscreen image
Up next

The future of elderly care

We place functional cookies to let this website function properly, analytic cookies to track usage and marketing cookies to improve your experience on this website. These cookies don’t use personal information. If you don’t want cookies you can change your cookie settings.

Cookie settings