Archilyse AG
Archilyse AG is a Swiss spatial data company that converts architectural floorplans into standardized, simulation-enriched datasets for AI training and building performance analysis.
Profile
Turn building floorplans into AI-ready datasets with simulated environmental metrics
Archilyse AG is a Swiss spatial data company that converts architectural floorplans into standardized, simulation-enriched datasets for AI training and building performance analysis. Founded by researchers addressing the 'garbage in, garbage out' problem in construction AI, the company has processed over 3,100 buildings comprising 45,000 apartments and 370,000 rooms through its open-source pipeline. Their datasets include Swiss Dwellings (CC BY 4.0) and Modified Swiss Dwellings (MSD), which add 367 simulation-derived metrics per area—from sunlight exposure to noise propagation—creating what they describe as 'context graphs' rather than isolated CAD files.
The core technology is released under AGPL license with commercial licensing available for proprietary forks. While not disclosing revenue, the company's value proposition centers on replacing inconsistent architectural drawings with verifiable spatial data that reduces model hallucination risks in proptech and urban planning applications.
Products by Archilyse AG
Who buys this
- Proptech startups training occupancy prediction models
- Municipal urban planning departments
- Academic researchers in computational architecture
- Real estate developers benchmarking portfolio performance
- AEC (Architecture, Engineering, Construction) software vendors
Strengths and what to watch
Strengths
- Open-core model with strong academic citations for dataset quality
- Simulation density (367 metrics/area) exceeds typical BIM exports
- Focus on Swiss market provides localized validation for EU regulations
Watch for
- AGPL licensing may limit enterprise adoption versus MIT/BSD alternatives
- Heavy reliance on Swiss data raises questions about global applicability
- No disclosed funding rounds or partnerships with major AEC players
Key Information
- Founded
- 2017
Frequently Asked Questions
What does Archilyse AG do?
Archilyse AG converts architectural floorplans into standardized datasets enriched with 367 simulation-derived metrics like sunlight exposure. Their AI-ready spatial data helps reduce model errors in proptech and urban planning, processing over 3,100 buildings through an open-source pipeline. (42 words)
How is Archilyse's building data different from CAD files?
Archilyse creates 'context graphs' with environmental simulations—like noise propagation metrics—rather than isolated CAD drawings. Their datasets include Swiss Dwellings (CC BY 4.0) and Modified Swiss Dwellings, adding 367 verified metrics per area for AI training. (42 words)
Who uses Archilyse's spatial data services?
Customers include proptech startups training occupancy models, municipal planners, academic researchers, real estate developers benchmarking portfolios, and AEC software vendors. The data's simulation density (367 metrics/area) suits applications needing verifiable building performance insights. (43 words)
Can I use Archilyse's datasets for commercial projects?
Core technology is AGPL-licensed, requiring open derivatives, but commercial licensing is available for proprietary forks. The Swiss Dwellings dataset uses CC BY 4.0, allowing reuse with attribution. Enterprises should evaluate AGPL's copyleft requirements. (45 words)
What are the limitations of Archilyse's approach?
AGPL licensing may deter some enterprises, and the focus on Swiss data raises questions about global applicability. While dataset quality is academically validated, there are no disclosed partnerships with major AEC firms or funding rounds. (44 words)
Where can I access Archilyse's building datasets?
Swiss Dwellings is on Zenodo (record 7788422), while Modified Swiss Dwellings technical details are on GitHub. The open-source pipeline documentation and licensing terms are available on Archilyse's website, with commercial options for proprietary use cases. (45 words)
Sources
- archilyse.standfest.science — Core product descriptions and dataset statistics
- zenodo.org — Swiss Dwellings dataset availability
- github.com — MSD dataset technical details
- archilyse.standfest.science — Pipeline architecture and licensing terms