Introducing Gaudi-1, our first SOTA proprietary model for automated constrained architectural generation

Through proprietary algorithms and specialized AI agents, Davis analyzes the full scope of a project — regulations, market data, and architectural inputs — in parallel.

Each dimension is retrieved, structured, and cross-referenced to build a coherent understanding of the site and its constraints.

Gaudi-1, our proprietary model, builds on this analysis to generate compliant architectural drawings that reflect these constraints from the outset.

These outputs are produced in seconds, then reviewed, refined, and validated by architects.

This approach transforms a fragmented, sequential process into a continuous system — where analysis directly informs generation.

Today, this enables fast, reliable feasibility studies and architectural plans.

Our long-term goal is to go further: to train a foundation model for architectural conception.