ifc-mcp
IFC MCP is a local MCP server that lets AI assistants inspect, create, and visualize IFC/BIM files.
Run Python in a Pyodide IFC runtime (IfcOpenShell, numpy, pandas, matplotlib, shapely) for inspection, validation, reports, exports, and generating new IFC files.
Open the local IFC viewer and get a stable viewer URL.
Show IFC files in the viewer by local path or generated download URL.
Clear the IFC viewer to remove all loaded models and BCF state.
Apply BCF/BCFZIP viewpoints to the viewer for issue-based navigation.
Return generated reports, BCF files, and IFC files as temporary local download links.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@ifc-mcp Open examples/sample.ifc and summarize the model."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
IFC MCP
IFC MCP lets your AI assistant inspect IFC/BIM files, create IFC files, create reports, open IFC models in a viewer, and apply BCF viewpoints.
How It Works
After installation, ask your AI client in normal language. The AI client starts IFC MCP when it needs the IFC tools and sends your request to the local server. IFC MCP reads IFC files locally, runs the needed IFC logic in a local Pyodide runtime with IfcOpenShell, and returns the answer to the chat.
![]()
Good example prompts:
Open examples/sample.ifc and summarize the model.Count the walls, doors, windows, spaces, and property sets in examples/sample.ifc.Find missing names, classifications, and suspicious property values in this IFC file.Create a CSV room-area report from this IFC file.Create a simple IFC file with one building, one storey, and four walls.Open the IFC viewer and show examples/sample.ifc.Apply C:\path\to\viewpoint.bcfzip in the open IFC viewer.Clear the IFC viewer.
If IFC MCP creates a report, BCF file, or new IFC file, it returns a local download link. If you ask to view a model, IFC MCP opens a local viewer link. If your AI client does not open that link automatically, open it in your browser.
Large IFC files can take more than a minute to process. Cancelling a request does not stop Python already running; wait for it to finish before trying again.
Related MCP server: IFCX MCP
Installation
1. Prerequisites
Install Node.js 20 or newer from the official Node.js download page. Choose the LTS version for your operating system, run the installer, then reopen the app where you want to use IFC MCP.
IFC MCP uses npx, which is included with Node.js. If your app later says
node or npx is not found, reinstall Node.js LTS and reopen the app.
2. Choose your client
ChatGPT desktop supports local MCP servers over STDIO.
Open the ChatGPT desktop app.
Open
Settings.Select
Plugins, thenMCPs.Click
Add. If prompted for a server type, chooseSTDIO.Enter:
Name:
ifc-mcpCommand to launch:
npxFirst argument:
-ySecond argument:
ifc-mcp
Leave the environment variable fields empty and save the server.
Restart the ChatGPT desktop app.
Start a new chat, type
/mcp, and confirm thatifc-mcpis connected.
For more detail, see OpenAI's MCP documentation.
Option 1: Use the Claude Desktop UI
Open Claude Desktop.
Open
Settings.Open
DeveloperorDesktop appdeveloper settings.Click
Edit Configif that button is available. This opens the Claude Desktop MCP config file.Paste the JSON below, save the file, then fully quit and reopen Claude Desktop.
{
"mcpServers": {
"ifc-mcp": {
"command": "npx",
"args": ["-y", "ifc-mcp"]
}
}
}If the file already contains JSON, add only the ifc-mcp entry inside the
existing mcpServers object. Do not create a second mcpServers object.
Option 2: Open the config file directly
Use this only if you do not see Edit Config in Claude Desktop:
Windows:
%APPDATA%\Claude\claude_desktop_config.jsonmacOS:
~/Library/Application Support/Claude/claude_desktop_config.json
Paste the same JSON from Option 1, save the file, then fully quit and reopen Claude Desktop.
Check
To check the setup, open Claude Desktop, start a new chat, click the + button
near the chat box, and look for Connectors. IFC MCP should appear there after
Claude Desktop has restarted. You should see the tools listed in the
How It Works section above.
For more detail, see Claude's local MCP server guide.
Option 1: Use the install button
This opens VS Code and adds IFC MCP.
If VS Code asks whether you trust the server, accept it. Reload VS Code if the tools do not appear. Open Chat and use the tools/configure-tools button to see the IFC MCP tools.
Option 2: Add .vscode/mcp.json manually
{
"servers": {
"IFC MCP": {
"type": "stdio",
"command": "npx",
"args": ["-y", "ifc-mcp"]
}
}
}Reload VS Code after saving the file.
Open Chat and use the tools/configure-tools button to see the IFC MCP tools.
For more detail, see the VS Code MCP server docs.
Option 1: Use the Codex settings UI
Open VS Code.
Open the Codex panel.
Open Codex settings.
Select
MCP servers.Click
Add server.Enter:
Name:
ifc-mcpCommand:
npxArguments:
-y ifc-mcp
Save the server.
Reload VS Code or start a new Codex session.
Check that the IFC MCP tools are available in the new Codex session.
Option 2: Use the terminal command
Open
Terminal -> New Terminalin VS Code.Paste and run:
codex mcp add ifc-mcp -- npx -y ifc-mcpReload VS Code or start a new Codex session.
Check that the IFC MCP tools are available in the new Codex session.
For more detail, see the Codex MCP docs.
Option 1: Use a project JSON file
Open your project in VS Code.
Create or edit
.mcp.jsonin the project root.Add this JSON and save the file:
{
"mcpServers": {
"ifc-mcp": {
"command": "npx",
"args": ["-y", "ifc-mcp"]
}
}
}Reopen Claude Code in VS Code.
Type
/mcpand approve the project server if Claude asks.Check that
ifc-mcpis connected.
Option 2: Use the terminal command for all projects
claude mcp add --transport stdio ifc-mcp --scope user -- npx -y ifc-mcpThen reopen Claude Code in VS Code, type /mcp, and check that ifc-mcp is
connected.
If the claude command is not found, install Claude Code first.
For more detail, see the Claude Code VS Code docs and Claude Code MCP docs.
Add this server entry to your MCP client config:
{
"mcpServers": {
"ifc-mcp": {
"command": "npx",
"args": ["-y", "ifc-mcp"]
}
}
}Restart the MCP client after saving the config.
For more detail, see the Model Context Protocol docs.
Use Your Own Viewer
If your app already has an IFC viewer, you can reuse IFC MCP's normal tool
interface and apply viewer actions in your app instead of opening the default
local viewer. The MCP client still calls tools like show-ifc-file and
set-bcf-view; your app only defines what those viewer tools do.
Provide handlers that receive local file paths. Replace the functions below with your app's viewer actions:
import { createServer } from "ifc-mcp";
import { serveStdio } from "@modelcontextprotocol/server/stdio";
serveStdio(() => createServer({
viewer: {
"open-ifc-viewer": () => openThisViewer(),
"show-ifc-file": ({ file_path }) => loadIfcIntoThisViewer(file_path),
"set-bcf-view": ({ bcf_path }) => applyBcfToThisViewer(bcf_path),
"clear-ifc-viewer": () => clearThisViewer(),
},
}));Return MCP tool results from your handlers: content contains feedback, and
isError: true indicates failure.
Pass files selected in your app and callbacks for your existing viewer.
Here, file is a selected IFC File, and viewer is your viewer instance:
import { createIfcMcpHost } from "ifc-mcp/browser";
const files = new Map([
[file.name, { name: file.name, source: file }],
]);
const host = createIfcMcpHost({
files,
viewer: {
"show-ifc-file": (name, bytes) => viewer.add(name, bytes),
"set-bcf-view": (name, bytes) => viewer.add(name, bytes),
"clear-ifc-viewer": () => viewer.clear(),
},
});
const result = await host.handleToolCall({
name: "show-ifc-file",
input: { file_path: file.name },
});File callbacks receive a filename and Uint8Array bytes. The BCF callback
must return an MCP tool result describing whether the viewpoint was applied.
Use host.tools for the tool definitions and host.getGeneratedFiles() for
generated downloads. Call host.clearFiles() when those downloads are no longer needed.
Local Development
For local development from this checkout:
npm install
npm testThe workspace includes .vscode/mcp.json, so VS Code can run the server from
this checkout.
Security
IFC MCP keeps the IFC work on your own machine. It reads your IFC files locally, runs the IFC analysis locally, and opens a local viewer. It does not publish the viewer to the public internet.
Generated reports, BCF files, and IFC files are temporary local download links. IFC MCP does not write generated output files into your project folder by default.
Use IFC MCP only with IFC/BIM files and project folders you trust. It can read the files you ask it to use, and incorrect or misleading model data can lead to wrong reports or generated IFC files.
Available Tools
5 toolsclear-ifc-viewerclear IFC viewerA
Clear the already-open local IFC viewer by removing all loaded IFC files and BCF viewpoint state. This keeps the same stable viewer URL and does not close the browser tab. Use show-ifc-file after this to load a fresh IFC model.
Inputs: no named parameters.
open-ifc-vieweropen IFC viewerA
Open or show the local IFC viewer and return its stable viewer URL. This tool does not load IFC files and does not change selection/isolation state. After this tool returns, use the returned URL as the viewer address. If the client can open links, open that local URL; otherwise tell the user to open the URL in a browser or webview. Use show-ifc-file to show IFC files. Use set-bcf-view to apply a BCF viewpoint file. Do not use browser automation to click or select elements in the viewer.
Inputs: no named parameters.
run-pythonrun PythonA
Execute LLM-created Python in the SDK Node/Pyodide IFC runtime. Use this tool for IFC inspection, validation, reports, exports, and creating new IFC files.
Runtime context: Pyodide uses the bundled Pyodide 0.28.2 runtime. The package set includes micropip, ifcopenshell, numpy, pandas, matplotlib, shapely, and typing-extensions. The Python code is passed directly to the SDK.
File inputs: pass IFC/IFCXML/IFCZIP file paths in the files array. In the desktop Node server these are local filesystem paths; in a browser host they are browser-defined paths resolved by that host. Generated Python can open those files by name, e.g. ifcopenshell.open("model.ifc") or open("model.ifc", "rb").
Return contract: this tool exposes stdout, stderr, saved_files, uploaded_files, and any result returned by the raw SDK. Generated files are returned as temporary localhost download URLs in saved_files. Do not assume files were written to the user's working directory.
Inputs: code (required), files, timeout_seconds.
set-bcf-viewset BCF viewA
Set the view state in the already-open local IFC viewer by applying a BCF/BCFZIP viewpoint file. Pass either a local BCF/BCFZIP file path or a generated BCF download URL from run Python saved_files. This tool does not show/load IFC files; use show-ifc-file for that. It should not open a new browser tab and should not use browser automation. To generate a BCFZIP with run-python, write a .bcfzip file, then call set-bcf-view with saved_files[0].url:
from uuid import uuid4
import ifcopenshell
from bcf.v3 import model as mdl
from bcf.v3.bcfxml import BcfXml
from bcf.v3.visinfo import VisualizationInfoHandler
model = ifcopenshell.open("sample.ifc")
gid = model.by_type("IfcWall")[0].GlobalId
component = mdl.Component(ifc_guid=gid)
visinfo = mdl.VisualizationInfo(
guid=str(uuid4()),
components=mdl.Components(
selection=mdl.ComponentSelection(component=[component]),
visibility=mdl.ComponentVisibility(
default_visibility=False,
exceptions=mdl.ComponentVisibilityExceptions(component=[component]),
),
),
)
bcf = BcfXml.create_new(project_name="IFC MCP")
topic = bcf.add_topic("Review wall", "Review wall", "ifc-mcp", topic_type="Issue", topic_status="Open")
topic.add_visinfo_handler(VisualizationInfoHandler(visinfo))
bcf_path = "view.bcfzip"
bcf.save(bcf_path)
print(str(bcf_path))Inputs: bcf_path (required).
show-ifc-fileshow IFC fileA
Show one IFC file in the already-open local IFC viewer. Pass either a local IFC/IFCXML/IFCZIP file path or a generated IFC download URL from run Python saved_files. Use this for IFC model display only. Repeated calls add multiple IFC files to the same viewer. Returns the same stable viewer URL; after this tool returns, use that URL as the viewer address. Use open-ifc-viewer to open the viewer and set-bcf-view to apply a BCF viewpoint file.
Inputs: file_path (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 5 tools
Each tool has a distinct purpose: opening the viewer, loading IFC files, clearing the viewer, applying BCF viewpoints, and executing Python code. There is no overlap in their core functions, and the descriptions clearly delineate the boundaries between them.
All tool names follow a consistent lowercase hyphenated verb-noun pattern (e.g., open-ifc-viewer, run-python, show-ifc-file). The verbs are distinct and the nouns are related to the domain, making the naming predictable and easy to navigate.
Five tools is well-scoped for an IFC viewer and scripting server. Each tool covers a necessary function without redundancy, and the count is within the ideal range for a focused MCP server.
The tool set covers the core viewer lifecycle (open, display, clear) plus BCF viewpoint application and a general-purpose Python execution tool for complex IFC operations. Minor gaps exist, such as no dedicated tool to remove a single loaded file or close the viewer, but these can be worked around with clear-ifc-viewer or run-python.
Maintenance
Related MCP Connectors
AI Hub for AEC — 50+ 3D formats, clash detection, ACC integration via Autodesk Platform Services.
Design buildings as IFC4 BIM models from a brief: an AI architect that plans, builds and edits.
Convert Revit files to XKT, IFC, or DWG and query BIM data via natural language.
Read, query, edit and render DWG drawings with AI. No AutoCAD required.
Related MCP Servers
- AlicenseAqualityDmaintenanceEnables AI assistants to interact with Autodesk Revit to query project data, manage elements, and execute generated code via the Model Context Protocol. It provides full compatibility with GitHub Copilot and Claude to automate BIM modeling workflows.13MIT
- AlicenseCqualityDmaintenanceEnables AI assistants to create, edit, and export IFC5/IFCX building information models through natural language, handling spatial structure, elements, geometry, metadata, validation, and export.7326MIT
- AlicenseAqualityDmaintenanceEnables AI agents to load, query, and analyze IFC building model files, including spatial structures, elements, properties, materials, and geometry.201MIT
- FlicenseAqualityDmaintenanceEnables AI assistants to query standardized building classifications, properties, and data dictionaries from buildingSMART for BIM model enrichment.93-