For architects and engineers working in BIM

Ask your model a question. Get a checked answer.

Aaron.ai reads your IFC in the browser and computes the numbers. An AI answers from them, five more check the answer, and every figure links back to its element.

How it works

  1. IFC4Your modelupload an .ifc
  2. ExtractionPsets · Qtos · unitsweb-ifc reads every element, in your browser
  3. TablesComputed by Aaron.aisizes, areas, openings per roomDoors 0.9 mRoom 12.4 m²
  4. AnswerConclusion + tableone AI, from the tables only
  5. CheckFive models reviewonly what they dispute shows⚠ Check this

Numbers from the file, not from an AI

Sizes, areas in m² and ft², openings per room, totals: Aaron.ai computes them from your IFC. The AI only reads those tables, so it can't make up a value.

An answer you can use

The conclusion first, then a comparison table per element: meets, doesn't meet, or check. Code questions (FHA, ADA, IBC, FBC…) name the requirement they apply.

Checked by five other models

GPT nano, Gemini Flash-Lite, DeepSeek, Qwen and GLM check every claim of the answer. You only hear from them when something needs a second look.

See the model in 3D

Open the IFC in the browser with its own colours, cut sections, filter by category, and click an element to ask about that element alone.

Projects keep everything

Several models per project (architecture, structure, MEP) and every question asked about each one, saved.

Cost repositories

Your own cost table, public ones and a community marketplace. Coming in the next releases.

A router, not a judge

Aaron.ai reads the file and does the arithmetic. It never gives a technical opinion of its own: the models answer and check each other, you see where every number comes from, and the decision stays with you.

Start a project