ML Conformer Generator

MLConformerGenerator is an advanced molecular generation platform for spatially-aware, structure-guided molecule discovery.

It enables researchers and engineers to design novel molecules that are optimised for 3D structural compatibility, binding affinity, and synthesis feasibility. By combining modern generative modelling, reinforcement learning, and fragment-based design strategies, it provides a unified API-driven and interactive workflow for next-generation drug discovery.

Core capabilities include shape-guided molecular generation for protein binding pockets and arbitrary 3D constraints, objective-driven optimisation using reinforcement learning with custom scoring functions, reference-based conformer similarity for scaffold hopping and ligand-based design, and fragment-based generation for controlled molecule editing and growth. Advanced features further enhance usability and performance through pharmacophore-constrained generation, low-strain conformer production, fragment library guidance for synthetic accessibility, high-throughput generation for rapid exploration, and fast expansion of molecular datasets and compound libraries.

The platform integrates seamlessly into existing cheminformatics and computational chemistry pipelines via a robust API, supporting both research and production environments. It also enables agentic workflows through MCP integration for iterative, closed-loop molecular optimisation. Users can interact through a web-based Molecule Designer with real-time 3D generation, as well as an Admin Dashboard for task management, system monitoring, and configuration. With low GPU cost, high throughput, and flexible deployment options, MLConformerGenerator accelerates scaffold discovery and improves hit quality across the drug discovery pipeline.