The AEC industry is at an inflection point. While AI has transformed nearly every other domain, from code generation to drug discovery, the built environment has remained stubbornly analog in its workflows.
The Problem
Most BIM tools today are monolithic, proprietary, and designed for a pre-AI world. They're powerful, but they weren't built to integrate with modern machine learning pipelines, natural language interfaces, or real-time collaboration at scale.
# Traditional clash detection: rigid, rule-based
def check_clashes(model_a, model_b):
for element_a in model_a.elements:
for element_b in model_b.elements:
if element_a.bbox.intersects(element_b.bbox):
yield Clash(element_a, element_b)This brute-force approach works, but it doesn't understand the building. It can't tell you that a duct clash near a structural beam is critical while one inside a ceiling void is probably fine.
Our Approach
At Pareidolia Labs, we're building tools that combine domain expertise with modern AI capabilities:
- Semantic understanding - Models that understand what building elements mean, not just where they are
- Free and accessible - Core tools are free. Open source where we can, always transparent
- Developer experience - APIs and interfaces designed for engineers, not just end users
"The best tool is the one you can inspect, modify, and trust."
What's Next
We're working on several fronts:
- Amorphous - Revit add-ins that actually respect your time. Batch operations, workflow automation, and solutions for the problems Autodesk ignores.
- Checkd - A modern alternative to Navisworks, built from the ground up with AI at its core. Clash detection and model review, without the bloat.
- Saccade - An AI agent for AECO. Connects to your tools, understands your models, and does real work across your entire stack.
Our Philosophy
The software behind our buildings shouldn't be a black box. We're for open data formats, open APIs, and tools you can actually inspect.
// AI-driven, context-aware clash analysis
const results = await analyzeClashes(model, {
context: "mechanical-structural",
severity: "auto",
suggestions: true,
});Stay tuned. We'll be sharing more about our technical architecture, design decisions, and lessons learned as we build in public.
Curious? See what we're building on GitHub.