Screenshot of a BIM Chat platform interface with sections for project management, building overview, export & reports, and hardware management. The building overview shows a hierarchy of floors and objects in a building, and there is a chat window asking about the number of doors on the first floor.

BIMChat

BIMChat is an advanced Building Information Modeling (BIM) platform that transforms how you interact with building data. Upload IFC files, chat with your building models using natural language, manage hardware catalogs, and generate comprehensive reports - all through an intuitive web interface.

Problem

Professionals across architecture, engineering, and construction rely on IFC files, but making sense of them often requires a patchwork of heavyweight tools, manual processes, and guesswork.

Solution

BIMChat lets you chat with your building models, run compliance checks, and generate reports, all from a lightweight web app that runs locally.

How

By combining modular microservices with structured IFC parsing and optional AI-powered interfaces, BIMChat delivers a fast, local-first experience that makes BIM data searchable, navigable, and actionable.

Screenshot of a BIM chat interface displaying building project data, including hierarchical building overview, file upload options, project details, export reports, hardware management, and a chat with a query about the number of doors on the first f
Screenshot of an Excel spreadsheet titled 'lfcFurnishingElement_Sofa Objects' with columns for Object ID, IFC ID, Global ID, Name, Description, Tag, Level/Floor, Space, Material, and other details, containing data about furniture objects.
A screenshot of a PDF document titled 'Building Composition Analysis' containing a bar chart and a pie chart. The bar chart shows object distribution by type with categories like dining tables, furniture, beds, and counters. The pie chart depicts bui
A digital report page titled 'Spatial Distribution Analysis' showing a bar chart comparing object distribution across different floors, with a legend indicating first floor and Level 1, and a table summarizing total objects and primary components for
Table titled Detailed Object Inventory listing various furniture objects, their parent types, counts, and sample object identifiers.

Deep-dive

Take a closer look at what went into delivering this solution.

Discovery

The point of creating BIMChat wasn’t to build a product. It was to answer the questions that traditional BIM software makes needlessly difficult:

  • What if I could just ask the model?

  • What if I could pull simple reports out of a lightweight app rather than loading the entire 3D model?

  • What if IFC data could drive real decisions, not just compliance paperwork?

The earliest prototype was simple: a parser, a hierarchy browser, and a chatbot. But it quickly became clear that there was value in joining the dots between parsing, understanding, validating, and presenting BIM data, all in one place.

Challenges

  • IFC Complexity

    IFC is a deep and messy standard. Pulling structured, queryable data from these files required advanced parsing, custom object mapping, and many edge-case workarounds.

  • Natural Language Alignment

    Mapping user questions (“How many toilets are there?”) to structured queries was difficult, especially when labels varied between projects. Normalising these disparate types allowed the system to infer intent from poorly worded queries, which meant building context-aware filters on top of the AI.

  • Local-First Deployment

    It was important that BIMChat could run locally for privacy and speed, but still offer modern features like LLM querying. That meant balancing containerised architecture with optional cloud-powered services.

  • Interface Simplicity

    BIM tools often assume a technically advanced user, while much of the data could be useful to those not trained in complicated modelling software. Creating an interface that was intuitive enough for a stakeholder but powerful enough for a BIM manager was key.

Use of AI

AI was the glue that brought the data to life. Using Claude Code to build the functionality brick by brick and integrating OpenAI’s API to power natural language queries.

Key use cases:

  • Translating human language into IFC-based object queries

  • Filtering out irrelevant data before answering

  • Generating structured, readable report outputs

The AI integration is opt-in, a critical decision for privacy-sensitive projects. And since BIMChat’s backend structures the data first, the models can focus on understanding questions rather than decoding broken inputs.

Why ikirugai?

ikirugai uniquely empowers agile SMEs to rapidly test and adopt the latest AI and automation technologies in practical, measurable ways. You can leverage your speed and adaptability to drive growth and efficiency right now, long before larger competitors have even left the planning room.

“This project demonstrated ikirugai’s ability to understand business needs, work pragmatically within constraints, and deliver a genuinely useful outcome. We’d happily recommend ikirugai to others exploring smart, lean ways of improving internal workflows.”

Toren Consulting

“I just had no idea that the problems I was facing could be so quickly solved; ikirugai somehow made it all feel easy, and now I understand the potential of what is possible with what I have.”

Global Workforce Placement Company

How?

Free 20-minute Discovery Call

Magnifying glass with overlapping circles design inside
Teal colored, curved arrow pointing right on a white background.

Tailored follow-up with next steps

Icon of two overlapping documents with a circuit brain design
Teal right arrow on black background

Prototyping or workshop within days

Illustration of a presentation board showing three bullet points and an analytics icon with arrows encircling a bar chart.

Let’s chat!

If you want to find out more, please fill out some info to arrange a FREE 20-minute discovery call!