what is roadrisk-ai-v1-0-0 fr?

RoadRisk AI is a browser-based tool for spotting road cracks and estimating pavement condition. It combines two pieces working together: a vision model that looks at photos of road surfaces to find cracks and surface damage, and a separate model that looks at structured data, such as current weather and traffic conditions, to judge how risky a given stretch of road is right now. The README describes this as an agent-based design, meaning the system is meant to pull together live weather feeds, traffic data, and camera images on its own rather than needing someone to manually combine those sources each time. The stated goal is to help infrastructure teams plan road maintenance and reduce how much manual inspection they need to do. There is no build process. You clone the repository and open the index.html file directly in a browser such as Chrome, Firefox, Edge, or Safari, since everything runs client side with no server required. From there you upload or point the tool at a JPEG or PNG photo of a road, and it detects cracks in the image while separately combining live weather and traffic inputs into a risk score. Settings for the weather API, traffic API, and the sensitivity of the risk score live in a config.json file that you edit directly. Running it needs an internet connection for the weather and traffic data, a recent version of one of the browsers listed above, and at least 4 GB of RAM if you are processing a lot of images at once. Risk scores refresh on a timer, every 15 minutes by default, and that interval can be changed in the configuration file. The project is released under the GNU GPL v3.0 license, a copyleft license that requires anyone distributing modified versions to also share their source code.