Architext is the world’s first semantic generation platform for Architecture. Using nothing more than plain language, users are albe to generate a rich variety of residential floorplans. This enables anyone to produce a nearly infinite set of creative designs, regardless of skill level or background.

Architext uses state-of-the-art language models, specifically finetuned for the task, to almost quickly generate functional designs. Generated layouts can be saved to all popular design formats, allowing it to simply integrate into existing workflows using products like Rhino, Revit, or AutoCAD.


Upcoming Work

Design Performance Evaluation

Architext can be augmented with further machine learning processes to perform design performance evaluations in an easy and intuitive way, bypassing traditional complexity



Theodore Galanos

Twitter  |  Linkedin

Theodore works at the intersection of Design and Intelligence, specializing in the development of computational design technologies for the built environment that seek to generate, extract, collect, and articulate Design Intelligence, bring together disciplines, and enable effective, efficient, and most of all, creative design processes. He is a Senior Researcher in the City Intelligence Lab of the Austrian Institute of Technology where he leads the development of InFraRed, the first AI-Driven, performance-based urban design tool.


Tyler Lastovich

Twitter  |  Linkedin

Tyler is a multidisciplinary engineer + strategist building products in deep tech, investing, and creative sectors. Previous experience covers a wide range of technical topics, including supercomputing at Cray Inc. and recently leading strategy at synthetic media startup Generated Photos. Having founded three companies previously, Tyler now chooses to push at the leading edges of technology, focusing primarily on natural language and quantitative investing.