Maxime Baudin

Baudin describes starting Modly while building a prototype that needed 3D assets. Hosted tools felt expensive and restrictive, so he made a local application and later opened its source. His sponsorship account identifies him as an independent developer focused on 3D, while his portfolio connects that work with web applications and games.[1][2]

His portfolio also presents Orvix, a semantic codebase-navigation CLI, and Crawlwise, a website audit product. These projects extend his interest in making existing technical capabilities easier to operate through focused tools.[1]

Connect open models to a desktop workflow

Modly runs image-to-3D generation on the user’s machine. Its extension system separates model integrations from the application, using manifests to define runtime entry points and downloadable weights. Workflows connect model and processing nodes. The application can also smooth and decimate imported meshes, keeping the result in its workspace.[3]

Give agents a way to operate the same application

A standard-library Python CLI can drive a running Modly instance through its local interface. This offers scripts and agents a route to the application’s workflow without manipulating the desktop UI. Model choice and hardware still determine what can run; the desktop interface is an integration of upstream models, not a claim to have trained them all.[3]