ChatGPT and Claude are the best general assistants there are, and Treform runs on the same models. But in a chat window they can only talk about your building, or hand you a script to debug. AI plugins for a single program put a chat box inside Rhino or Revit, and stop at the edge of that app.
Treform works differently. One agent runs inside Rhino, Grasshopper, SketchUp, Revit, Blender and AutoCAD at the same time, carries architectural data and the Handbook with it, and has tools that actually run: rendering, 2D drawings, laser-cut templates and the ones you build yourself.
At a glance
"Other plugins" means AI assistants built for one program, usually Rhino or Revit. Where products differ we say "varies" rather than guess; comparison based on publicly available information as of October 2026.
The core difference: a chat window vs an agent in your software
Ask ChatGPT or Claude for a building and you get a description, or a script you paste into your software and debug yourself. The model is capable; it just has no hands. Every change means another prompt, another paste, another fix.
Treform is the hands. It sits inside the software you already use, reads the model that's open, and makes the edit there. You describe the change in a sentence, or drop in a reference photo, and the model updates in place. The same models do the thinking; the difference is that the result lands in your file.
One app vs your whole stack
Single-app plugins are useful inside their program and useless outside it. A Rhino assistant can't document in Revit; a Revit assistant can't mass in Rhino or render in Blender. The work still moves between apps by hand, through exports and copy-paste.
Treform runs one command across all of them. Mass in Rhino, detail it, document in Revit, drop it into SketchUp or Blender for a render, and it's the same agent and the same model the whole way. No file exports, no re-describing the project each time.
Architectural judgment, not just geometry
A general model knows roughly what a building is. It doesn't know what a program needs, where structure goes, or what a code requires, so you check every output. Most plugins are a prompt wrapped around that same model.
Treform ships with built-in architectural data and the Handbook: program, structure and judgment applied on every edit. That's why its models come out clean and complete, built with modeling logic rather than generated geometry you tidy up afterwards.
Models, tools and training
Every Claude and GPT model is available in Treform, switchable per task. Preset tools cover rendering, /2d drawings, massing-to-detail, styles, site and laser-cut templates, and you can turn your own workflows into reusable tools. Train the agent on your projects and standards and it models the way your studio does: your layers, families, naming and detailing habits.
Pricing compared
ChatGPT and Claude subscriptions cover their chat apps; none of it runs inside your design software. Single-app plugins are priced per app, so a two-app studio pays twice.
Treform starts free, with 3 AI edits a month and no card. Pro is $39/mo billed yearly ($49 monthly) for 30 edits; Studio is $124/mo billed yearly ($149 monthly) for 120 edits and 100 renders. Bring your own Claude or OpenAI key on BYOK, $19 a month, with no edit cap. Teams are $40 per seat per month, 10 seats minimum. Full pricing →
Which should you pick?
If you're serious about your modeling and the quality of your models, and you work in architecture or design, Treform is the one for you. It's free to start.