PyNET Viewer — BIM for VS Code - Visual Studio Marketplace

PyNET Viewer

PyNET Viewer brings federated BIM models into VS Code and connects your AI assistant to Autodesk Navisworks, Revit and Civil 3D — so it can explore, explain and operate the model in natural language, without leaving the editor.

A federated model open in the PyNET Viewer, with element properties and the recent-models list

📥 Request your 30-day Trial here:

Contact: info@raendt.com to request access.

🧊 What this extension does


3D BIM Viewer	Open a `.pnt` package and navigate the federated model — spatial tree, element properties, sections, measurements and clash results. Free, and works on its own.
MCP Bridge setup	One click installs `pynet-mcp-bridge` and configures every AI client it finds: Claude Desktop, Claude Code, Codex, Cline, Roo Code and GitHub Copilot.
Reference docs	Installs the routing guides for each Autodesk host locally, so your assistant knows which API to reach for before it writes a line.

💬 What you can ask

Once the bridge is wired, the assistant talks to the model and to the viewer at the same time — so it can point at what it is describing:

"Which clashes involve the HVAC model? Show me the worst one."

"Isolate the structural discipline and hide everything else."

"List every duct type on level 2 with its element id."

The assistant breaking down 3,706 clashes by test, with the selected clash pair highlighted in the viewer — column in red against slab in green

_{The assistant reads the package, explains what it found, and highlights the
pair it is talking about — column in red, slab in green, everything else hidden.}

📋 Requirements

Requirement	Detail
Windows	The Autodesk hosts are Windows-only.
Python 3.10+	Mandatory — both the bridge and the viewer server run on it. Auto-detected on `PATH`, or set `pynet.pythonPath`.
PyNet plugin	Only needed to drive Autodesk from your AI client. The viewer works without it.

🚀 Getting started

Install the extension. On first activation it installs the bridge and configures every AI client it detects — no action needed. (Turn it off with pynet.autoInstallBridgeOnStartup.)
Restart your AI clients so they pick up the new MCP server.
Open a model. Click the PyNET icon in the activity bar → Open BIM Model, and pick a .pnt package. Everything you open is kept in a recent list.

If a client was closed during setup, run PyNET: Install / Repair MCP Bridge from the command palette at any time — it is safe to re-run and re-detects everything.

📚 The reference library

Writing a script for Navisworks, Revit or Civil 3D means knowing which API applies, which imports are allowed, and how the host expects to be addressed. The extension installs those routing guides to %APPDATA%\Pynet\Library and offers to open them the first time.

Open that folder as a workspace and your assistant picks the routing up automatically — that is what turns "write me a clash report" into a script that runs first time. Reopen it whenever you like with PyNET: Open Reference Library.

⌨️ Commands

Command	Action
PyNET: Open BIM Model	Opens a `.pnt` package in the embedded 3D viewer.
PyNET: Install / Repair MCP Bridge	Installs or upgrades the bridge and reconfigures every detected AI client.
PyNET: Open Reference Library	Opens the installed reference docs as a folder.
PyNET: Clear Recent Models	Empties the recent models list.

⚙️ Settings

Setting	Default	Description
`pynet.pythonPath`	`""`	Python 3.10+ executable. Empty = auto-detect (`python`, `py -3`).
`pynet.viewerPort`	`0`	Local viewer server port. `0` = pick a free one.
`pynet.autoInstallBridgeOnStartup`	`true`	Install the bridge and configure clients on first run.

🤖 How it works

The viewer is a That Open engine build served by a bundled local server and embedded in a VS Code webview. A control channel lets the MCP bridge drive the scene — select, isolate, fit, highlight clashes, read properties — so your assistant can point at what it is talking about while it talks about it.

Important note on AI providers: PyNET is the integration layer between AI models and Autodesk tools. Access to AI models (Claude, OpenAI, etc.) is not included — you bring your own client and subscription.

🔒 Privacy & Security

Local execution. The viewer server binds to localhost; the model never leaves your machine.
Nothing uploaded. No BIM data is transmitted externally at runtime.
Validated scripts. Anything the AI sends to an Autodesk host passes a static validator first.
AI-generated code may be processed by the AI provider you choose to use.

🧩 The PyNET Ecosystem

Component	Repository	Purpose
PyNet Platform	PyNet	Navisworks/Revit/Civil 3D plugin hosting the Python.NET engine
PyNet Bridge (MCP)	PyNetBridge	MCP server connecting AI models to PyNET, with secure script validation
PyNet Library	PyNetLibrary	Reference scripts and AI context for the three hosts
PyNet for VS Code	PyNetVSCode	This extension — embedded viewer + one-click bridge setup

❓ FAQs & Support

Questions about installation, configuration or usage? See the PyNET FAQs, or report an issue.

⚠️ Disclaimer

PyNET Platform is intended for professional use in BIM automation. Users are responsible for reviewing, validating and ensuring the correctness of AI-generated scripts before applying them in production environments.

PyNET Viewer — BIM for VS Code

RAENDT