InfoBIM Core

  _____        __      ____ _____ __  __
 |_   _|      / _|    |  _ \_   _|  \/  |
   | |  _ __ | |_ ___ | |_) || | | \  / |
   | | | '_ \|  _/ _ \|  _ < | | | |\/| |
  _| |_| | | | || (_) | |_) || |_| |  | |
 |_____|_| |_|_| \___/|____/_____|_|  |_|

The Capability Engine for BIM Automation

Turn your BIM scripts into standardized, agent-ready capabilities.

InfoBIM IFC is a modular runtime environment for Engineering Automation. It decouples the execution logic (how to run scripts, handle dependencies, manage Docker) from the business logic (validating pipes, counting doors, auditing models).

Whether you are a human running commands in the terminal or an AI Agent planning a complex workflow, InfoBIM provides the standard interface to interact with BIM data.

Overview

• CLI: infobim | check | run | list
• Engines: supports venv and colab via OntoBDC configuration
• IFC Support: built on top of IfcOpenShell for openBIM workflows

🚀 Quick Start

1. Install

InfoBIM is designed to require as little setup as possible to get started. You can currently use InfoBIM locally (Python venv) or in Google Colab, depending on your needs.

Local Installation

Local installation requires Python 3.10+ with pip and venv available. To create and activate a virtual environment, go to the project directory and run:

python -m venv .venv
source .venv/bin/activate

Then:

pip install infobim

To verify the installation, run:

infobim --version

And, welcome to InfoBIM world!

Google Colab Mode (Optimized for Notebooks)

All current InfoBIM features are available in Colab. You can check the example notebook here:

Google Colab Notebook

2. Run a Capability

Capabilities are the core of InfoBIM. They define the actions that can be performed on IFC models. When you execute a capability, InfoBIM will:

• Check whether the capability is available in the current environment.
• Execute the capability with the provided parameters.
• Return the execution results.

For this reason, only capabilities whose required parameters are satisfied are shown in the runtime menu. When you run:

infobim run

Interactive mode (terminal menu) shows a UI to select an available capability. Most InfoBIM capabilities require an IFC file path, so you will usually pass --ifc-path when running by ID.

3. List available capabilities

Lists all capabilities discovered in the current environment (after initialization and checks).

infobim list

4. System Health & Maintenance

InfoBIM includes self-diagnostic tools to ensure your environment is correctly configured.

Run System Check: Verifies if all dependencies (Python, Git, Venv, Docker) and project structures are in place.

./infobim check

Auto-Repair: If a check fails, you can try the repair mode, which attempts to fix common issues automatically (e.g., recreating venv, installing missing requirements).

./infobim check --repair

🧩 IFC Capabilities

A Capability is the atomic unit of work in InfoBIM. It wraps your Python scripts with:

1. Metadata: Name, version, description.
2. Schema: Formal definition of Inputs (arguments) and Outputs (return data).
3. Docs: Embedded documentation for both humans and LLMs.

Thus, InfoBIM IFC acts as a lightweight runtime layer for BIM automation, turning scripts into discoverable and repeatable capabilities.

• Runtime: Handles the environment (Python, Docker, IFCOpenShell).
• Registry: Discovers and registers available Capabilities.
• Interfaces:
  • CLI: Direct execution (./infobim run --id <capability_id> --ifc-path ./path/to/model.ifc).
  • TUI: Interactive Text User Interface (./infobim run --ifc-path ./path/to/model.ifc).

Examples:

infobim run --id org.infobim.domain.ifc.capability.list_buildings --ifc-path ./path/to/model.ifc
infobim run --id org.infobim.domain.ifc.capability.list_property_sets --ifc-path ./path/to/model.ifc --global-id '2T$yW1JQv9FQ0kqP8m9qXb'
infobim run --id org.infobim.domain.ifc.capability.inspect_element --ifc-path ./path/to/model.ifc --global-id '2T$yW1JQv9FQ0kqP8m9qXb'

🤖 For AI Agents

InfoBIM is Agent-First.

• Discovery: infobim list --json provides the tool definitions (compatible with OpenAI/Claude function calling).
• Deterministic Execution: Agents don't "guess" geometry; they call Capabilities that return precise data.
• Structured Output: All capabilities return strict JSON data, making it easy for LLMs to reason about the results.

📄 License

This project is part of the InfoBIM Community. Licensed under Apache 2.0.