Physics engine. Conversational interface.

An AI agent for structural engineers that removes tool friction, not engineering judgment.

Book a Demo Try Polaris

How Polaris Works

Polaris uses tools and skills, like a human engineer would.

Step 1

Structural Geometry

Import IFC or pull from organizational context (RAG).

Step 2

Load Analysis

Imposed, wind and per-zone loads, cascaded top-down through the structure.

Step 3

FEA Simulation

Cloud-based finite element analysis on all structural elements in minutes.

Step 4

Compliance Check

Per-element checks flagging under-designed, over-designed and deflection failures.

Step 5

Optimization

Size, material and other changes, converging on safety, deflection and more.

Step 6

Calculation Reports

PDF, XLSX and IFC exports — load cases, utilization ratios, code references.

Steps 3–5 run in an iterative optimization loop until targets are met

Engineer in the Loop

Polaris doesn't replace engineering judgment. It orchestrates the iteration-heavy work that sits around it.

You retain
  • Modelling assumptions and boundary-condition choices
  • Load-path interpretation and code-clause application
  • Safety judgment and professional sign-off (CEng, MIStructE, equivalent)
  • Design decisions between the alternatives Polaris proposes
Polaris handles
  • Meshing, solving and running FEA iterations on CalculiX
  • Cross-referencing every element against Eurocode 2/3 (and 7 for geotechnical work)
  • Propagating a model change through analysis, compliance checks and calculation reports
  • Assembling the auditable PDF with load cases, utilization ratios and code references

Who It's For

From small firms to enterprise engineering teams.

Structural Engineering Teams

Shorten iteration cycles from hours to minutes

  • Run full FEA in the cloud — no local compute bottleneck
  • Iterate sections, materials and reinforcement interactively
  • Eurocode compliance checks built into every analysis run
  • Calculation-ready reports for regulatory submission
Design-Build / EPC Firms

Optimize structural cost before you submit the bid

  • Test design alternatives quickly during the bidding phase
  • Reduce steel tonnage and concrete volume with iterative optimization
  • Re-run full analysis in minutes when parameters change
  • Auditable calculation reports for every design decision

Why Polaris

⊘

No More Rebuilds

Turn architect models into analysis-ready structural models automatically.

⟳

Reuse Past Projects

Find any model in seconds using natural language. Open in 3D and keep working.

⏱

Minutes, Not Hours

Cloud-based FEA runs in minutes. No waiting on local machines.

⚙

Iterative Optimization

Automated design-analysis loops converge on safety and cost targets.

✓

Eurocode Built In

EN 1990/1991/1992/1993 compliance checked on every run.

▤

IFC Native

Works with your existing models. Import from Revit, Tekla, or BlenderBIM.

☁

Cloud-Native

Scales on demand. Run from any device, no installation required.

▣

Submit-Ready Reports

Formatted for regulator and client review, with every number traceable to its source.

Frequently Asked Questions

How do I begin?

Just upload your IFC model. The agent will guide you through the process, from specifying load parameters to running a complete finite element analysis.

Does AI perform the actual calculations?

No. Polaris does not use AI to guess physics. AI acts as an orchestration layer to extract design intent. The actual math relies on a deterministic Eurocode cascade, the finite element analysis (FEA) runs on proven open-source solvers like CalculiX and Gmsh. The engineer remains in the loop to confirm parameters and validate the results.

What design standards and formats does Polaris support?

Polaris works natively with IFC models and calculates load combinations and serviceability checks against Eurocode standards (EN 1990-1999). Results can be exported as structured PDF reports, XLSX tables and optimized IFC files.

Can I use my own AI agent?

Yes. Polaris provides a standard Model Context Protocol (MCP) server. You can connect Claude, ChatGPT, or your own custom MCP-compatible agent to utilize visualization, IFC modification and analysis capabilities.

Is my project data used to train AI models?

No, the data remains under your control and is never used to train or fine-tune models. For strict data sovereignty requirements, we also offer private cloud (VPC) and on-premises deployments.

Is Polaris tied to a specific language model?

No, Polaris is LLM-agnostic. The platform's modular architecture means you aren't locked into a single provider, allowing you to leverage the language model that best fits your organization's capabilities and privacy requirements.

Do I need a powerful workstation to run it?

Not at all. Polaris is device-agnostic and cloud-native. All the heavy computation, including finite element analysis (FEA) and 3D processing, is handled in the cloud, allowing you to interact with massive structural models from any device with a standard web browser.

Backed by Antler