Deep Model Processing
vyzn’s core technology. DMP is the proprietary technology that transforms digital building models (BIM) into clean, structured, simulation-ready assets. Automatically, in minutes.
BIM models are built for documentation, not for performance analysis and simulation.
Turning a BIM model into a simulation-ready energy model requires fixing missing room definitions, incomplete geometry, and semantic gaps. Today, this translation is often performed manually by specialists on every project - a process that is slow, costly, and error-prone.
This bottleneck stands at the entry point of every energy simulation, every carbon calculation, and every performance assessment. Building analyses therefore happen too late and too rarely - and can barely contribute to better planning decisions.
Automated model intelligence at scale.
DMP ingests any 3D building model and outputs a validated, semantically enriched, simulation-ready reference model: the shared foundation for every analysis vyzn runs. No manual rework. No special authoring requirements from the architect or engineer.
All proprietary, all compounding.
Explore a processed model.
After DMP, every room, surface and zone becomes explorable. Orbit the building, peel back the floors, and inspect the semantic structure the engine reconstructs.
Principles
Universal preparation
Any 3D model becomes analysable, with no special authoring guidelines required.
Automatic enrichment
Missing values are filled in, overridable by the user.
Automatic error correction
Detected defects are fixed, transparently and adjustably.
Weeks of preparation compressed into minutes.
What previously required weeks of model preparation now happens automatically. That unlocks something structurally new: the ability to run dozens of variants (energy, emissions, comfort, cost) in the time a single traditional analysis used to take.
For building owners and planners, that’s a step-change in how early-stage decisions get made. For vyzn, it’s what makes the platform scalable.
A technical asset that gets stronger with use.
Proprietary pipeline
Built on COMPAS (ETH Zürich’s open-source computational design framework) - extended with proprietary algorithms and training data grown from hundreds of real-world building projects.
Data flywheel
Every building model processed improves the AI enrichment layer: envelope detection, element classification, room type prediction.
Automation advantage
Minutes, not weeks. A fraction of the cost. No manual effort.
Platform foundation
One cleanly processed DMP model feeds every analysis bundle: energy, emissions, comfort, cost, safety. The labour-intensive model preparation happens once, not once per use case.
Want to go deeper?
We’re happy to walk investors and partners through how DMP works under the hood.