Read the model
Ask questions in plain language — what's in this model, what changed, where does it fall short of the brief.
Revit AI, on your hardware
Run Revit AI on hardware you control.
Revit AI is powerful — and it belongs on your workstation. Revit Agents keeps the LLM, the MCP server and your project model local, behind a secure connector, while a governed cloud console decides what the AI is allowed to do.
Set company policy once, approve prompts per project, preview every change before it reaches Revit, and keep an auditable record of what ran and who authorised it.
See the full governance storyModels and project data never leave your workstation.
Local-first AI stack
The browser never talks to Revit. The AI runs where your data lives.
- 1Authenticated Revit Agents console
- 2Server-side policy and approval checks
- 3Secure connector to a paired workstation
- 4Local LLM and agent runtime
- 5Revit MCP server and Autodesk Revit
Capability is detected per workstation — the console reports what is genuinely reachable rather than assuming a live connection.
What is Revit AI?
Revit AI means large language models operating directly on your Revit model — reading it, reasoning about it and acting through the Model Context Protocol (MCP). The real question for a practice isn't whether to try it. It's where the AI runs, and who governs what it's allowed to do.
Propose, then preview
AI suggestions are analysed and previewed read-only before anything is written to Revit.
Automate the repetitive
Approved prompts turn recurring tasks into repeatable runs that come back verified and recorded.
Why Revit AI should run on your workstation, not a vendor's cloud.
Project models are commercially sensitive. Client data, design IP and delivery standards shouldn't be uploaded to a third party to get AI assistance. Revit Agents gives you the AI — and keeps the data local.
Cloud AI plugins
Convenient — at a cost you can't fully see.
- The project model is uploaded to a vendor's cloud
- Client data and design IP are handled off-site
- Ad hoc prompts that are hard to review or audit
- Little visibility into what the AI changed, and why
With Revit Agents
Local execution, central oversight.
- LLM, MCP server and model stay on the workstation
- A secure connector — the browser never talks to Revit directly
- Versioned, approved prompts with role-based access
- Analysis and preview before apply, plus an auditable run history
From approved prompt to verified Revit result
Every stage is explicit. Nothing is written to Revit until an authorised run is approved, and the result comes back to be verified and recorded.
- 01
Define policy
Company standards, permitted tools and safety rules.
- 02
Approve the prompt
A prompt version moves from draft to active.
- 03
Deploy to the project
Scoped configuration for that job's standards.
- 04
Analyse & preview
Read-only inspection of the proposed change.
- 05
Authorised run
The Revit write happens only after approval.
- 06
Verify & audit
A verified result, recorded in run history.
Write gate
Analysis and preview are read-only. The Revit write only happens on an authorised run, and returns a verified result to the audit record.
What teams are doing with Revit AI
Early access is built around the workflows BIM teams already know — performed by a governed agent instead of by hand.
Model checking
Compare the model against project standards and the brief before sign-off.
Documentation
Generate schedules and reports from the model with approved prompts.
Classification & cleanup
Find and correct inconsistencies across elements and data.
Drafting support
Let the agent propose edits for review instead of typing them by hand.
Bring governed Revit AI to your practice.
Early access is opening to a limited number of practices. Join the waitlist and we will be in touch as places become available.
Want the full picture? See how Revit Agents governs automation practice-wide