Built a Revit Chatbot with Just One AI Prompt!
AI inside Revit is no longer a future concept. In this tutorial, I show how to build a fully working Revit chatbot plugin using GitHub Copilot inside Visual Studio. The result is a ChatGPT-like experience running directly inside Revit, powered by .Net 8 and your own OpenAI API key.
This blog post explains what you will learn, why this workflow matters, and how AI can accelerate real Revit plugin development without sacrificing control.
Resources
What This Video Is About
This tutorial focuses on AI-assisted development rather than blindly generating code. The goal is not just to build a chatbot, but to show how structured prompts can guide AI to behave like a disciplined junior developer.
In the video, you will see how to
• Use GitHub Copilot Chat inside full Visual Studio
• Generate a complete Revit plugin from a single detailed prompt
• Build a ChatGPT-style chatbot UI using WPF
• Follow proper Revit API rules and deployment practices
• Iterate and fix bugs using follow-up AI prompts
Everything runs inside Revit, not a browser or external app.
Tools Used in This Tutorial
Before writing any code, the video walks through the exact tools required for this workflow.
You will need:
• Visual Studio Community Edition
• .Net desktop development workload
• GitHub Copilot with Chat enabled
• An OpenAI developer account and API key
Visual Studio Community is completely free and more than sufficient for professional Revit plugin development.
Designing the Revit Chatbot Plugin
A big part of this tutorial is the prompt itself. Instead of letting the AI guess, the video shows how to clearly define structure, architecture, and constraints.
The plugin is designed to:
• Run in Revit using .Net 8
• Appear as a new ribbon tab in Revit
• Open a modeless WPF chatbot window
• Store chat history and user input cleanly
• Include a separate settings window for API keys
Clear requirements lead to predictable AI output.
Revit API and Architecture Rules
The video emphasizes that Revit plugins have strict rules, and AI must be reminded of them.
The prompt enforces
• Correct use of the Addins deployment folder
• Use of .addin files and DLLs
• Implementation of IExternalApplication
• Implementation of IExternalCommand
• MVVM architecture with one class per file
This results in a project that looks and feels like a real production plugin.
Debugging and Running the Plugin
Once the AI finishes generating the project, the video shows how to run and test it properly.
You will see how:
• Visual Studio is configured to launch Revit automatically
• The plugin ribbon tab appears correctly
• The chatbot window opens as modeless
• The OpenAI API key is entered securely at runtime
• Chat responses work as expected
The API key is never saved to disk and only lives in memory for the current session.
Fixing Bugs With Follow-Up Prompts
AI-generated code is fast, but it still needs validation. The tutorial includes a real usability bug where chat text cannot be selected or copied.
Instead of fixing it manually, the bug is resolved by
• Writing a simple follow-up prompt
• Letting the AI update the WPF XAML and ViewModels
• Rebuilding and retesting inside Revit
This shows how AI can be used iteratively, not just once.
Why This Chatbot Does Not Modify the Model
One important design decision covered in the video is safety.
This chatbot is intentionally limited to guidance and explanations. It does not directly modify the Revit model.
Benefits of this approach include:
• Reduced risk of accidental model changes
• Clear separation between advice and automation
• A safer entry point for AI inside production projects
Model-modifying AI workflows require deeper architectural considerations and are covered in a separate video on the channel.
Who This Tutorial Is For
This video is ideal if you are
• Exploring AI inside Revit
• Learning Revit API development
• Interested in GitHub Copilot beyond VS Code
• Building internal BIM tools
• Curious about ChatGPT-style assistants in Revit
Whether you are new to plugins or already experienced, this workflow scales well as your tools grow.
Watch the Full Tutorial
To see the full process, including the exact prompt, project structure, debugging, and final chatbot in action, watch the embedded video above.
If you enjoyed this tutorial, consider subscribing to the channel. I regularly share Revit, Dynamo, automation, and AI-focused content to help you work faster and smarter 😊
