Table of Contents Show
AI generators now sit in almost every stage of an architecture project, from the first mood image to the feasibility study for a site. They do very different jobs, though: some produce pictures, some write text, and some generate floor plans or site layouts from rules. This list covers nine of them, what each one actually does, and how an architecture student can use it without handing the design thinking over to the software.
Short answer for students: start with one image generator (Midjourney or the image tool inside ChatGPT) for mood boards and early atmosphere studies, and use ChatGPT for writing and research support. Leave the planning tools, such as Finch, Autodesk Forma, Archistar and TestFit, for studio briefs about housing, density or urban sites, where your school may already provide access. For a wider list of apps, see our guide to AI apps for architects.
| Tool | What it generates | Best student use |
|---|---|---|
| MidJourney | Images from text prompts | Mood boards, material and light studies |
| ChatGPT | Text, plus images through its built-in image tool | Brief analysis, precedent research, editing project text |
| DALL-E | Images from text (now replaced inside ChatGPT) | Quick concept images if you already use OpenAI tools |
| DeepArt | Style transfer and AI images | Graphic experiments for presentation boards |
| Finch | Floor plans and layouts from rules | Testing unit layouts in housing studios |
| Archistar | Code and zoning compliance checks | Understanding how plans are reviewed for permits |
| Autodesk Forma | Massing options with sun, noise and wind analysis | Early site and massing studies |
| ReRender AI | Renders from sketches and model screenshots | Fast visuals for desk crits |
| TestFit | Site plans, parking and feasibility options | Real estate and urban design projects |
1. MidJourney
MidJourney is an image generator that turns written prompts into pictures. You use it through its Discord bot or its own website. Architects use it to test atmosphere, light and material ideas quickly, long before anything is modelled. It does not produce measured drawings: the plans and sections it shows are pictures, not geometry you can build from.
For students: use it for mood boards and early atmosphere studies, keep your prompts in a project log, and label every Midjourney image as AI-generated on your boards. Our Midjourney vs DALL-E comparison shows how the two handle architectural prompts.

2. ChatGPT
ChatGPT by OpenAI is a chat assistant that architects use to draft project descriptions, summarise briefs, structure reports and prepare client proposals. It can also produce images through its built-in image tool. Treat its answers as a first draft: it can state wrong facts with confidence, so check building codes, dates and references against the original source.
For students: ask it to explain a brief back to you, list questions for a site visit, or tighten the text of your project statement after you have written it yourself. Our guide on how architecture students should use ChatGPT gives example prompts.
3. DALL-E
DALL-E, another innovation from OpenAI, creates stunning visual representations from textual descriptions. Architects can describe a scene or a design concept, and DALL-E will generate a corresponding image. This tool is particularly useful for creating mood boards, conceptual visuals, and exploring design alternatives early. Note the update: inside ChatGPT, DALL-E 3 was replaced by OpenAI’s GPT Image model in March 2025, as recorded on the DALL-E Wikipedia page. If you generate images in ChatGPT today, you are using that newer model.
For students: it suits quick concept images when you already work in ChatGPT. Ask for several variations of the same prompt and compare them, rather than accepting the first image.
4. DeepArt
DeepArt uses neural networks to turn photographs into artworks in the style of famous artists. The product family has grown since this list was first written: its site now offers Deep Art Studio for generating images and videos from a text prompt, Deep Art Creator for image generation offline on a PC, and the original Deep Art Effects style filters. Architects can use it to see how a painterly or graphic style changes the reading of a render or site photo.
For students: try it on your own site photos or model shots to develop a consistent graphic style for a presentation board, then keep the original image next to the stylised one in your process work.

5. Finch
Finch is a generative design platform that helps architects create building layouts from a set of rules. It generates buildable floor plans based on a firm’s own standards and local codes, tests unit distribution and density across a site, and keeps area calculations updated as the design changes. Architects can compare several layout options early and carry the chosen one forward as BIM geometry, which also helps sustainable design decisions about density and layout get made at the right stage.
For students: in a housing studio, use it to compare unit mixes and layouts against your own hand-drawn options, and explain in your crit why you kept or rejected what it generated.
6. Archistar
Archistar started as an AI platform for site analysis and development design. Its focus has since moved to plan review: its main product, AI PreCheck, checks building designs for code and zoning compliance and is aimed at local governments, with a version for designers and certifiers who want to check plans before they submit. The company names Los Angeles and Austin among the cities using it.
For students: you are unlikely to need a licence, but its public material is a useful way to understand what a permit reviewer checks, which helps when you prepare a planning or code study for a design project.
7. Autodesk Forma (formerly Spacemaker)
Autodesk Forma is the AI-supported early-stage planning tool that replaced Spacemaker in May 2023, after Autodesk acquired Spacemaker in 2020. It helps architects and urban planners test massing options in complex urban environments and gives quick insights into sunlight, noise, wind and other environmental factors. These analyses help architects make informed design decisions at the stage when changes are cheapest.
For students: use it in urban or site-heavy studios to test two or three massing options against sunlight and wind before you commit to a scheme, and check whether your school’s Autodesk education access covers it.
8. ReRender AI
ReRender AI turns sketches and plain model screenshots from SketchUp, Revit or Rhino into photorealistic renders in the browser. It fills the role that NVIDIA’s GauGAN sketch-to-image demo once played, and that demo no longer appears in NVIDIA’s AI Playground. ReRender AI is useful for quick visualizations during the brainstorming phase; see our ReRender AI review for its speed, styles and limits.
For students: it is handy for a quick visual of a SketchUp or Rhino model before a desk crit. Keep the raw screenshot alongside the render so your tutor can see what came from your model and what came from the AI.

9. TestFit
TestFit is a generative design tool for site feasibility. Its Site Solver generates building and site configurations from your constraints, and its Parking Solver lays out surface and structured parking. Architects can input unit mixes, setbacks and parking targets, then compare the options TestFit generates, which makes early feasibility studies much faster.
For students: TestFit runs a programme for universities for use across a class or on specific projects, and its exports to Revit, SketchUp, AutoCAD (.dxf) and Excel (.csv) let you carry a feasibility study into your design model. It suits real estate, housing and urban design projects where cost and yield matter.
How to Use AI Generators in Studio Without Losing Marks
Schools set their own rules on AI tools, and those rules can differ between universities and even between studios on the same course. A few habits keep you on the safe side:
- Read your course or studio policy on AI before the first submission, and ask your tutor when it is unclear.
- Label every AI-generated image, render or text on your boards and in your portfolio.
- Keep a process log: the prompt, the date, the tool, and what you changed afterwards.
- Never present an AI image as a measured drawing. Plans, sections and details still come from your own model or drawings.
- Check every fact, code clause and reference that a chat assistant gives you against the original source.
For a student-focused shortlist with more tools, see our guide to AI tools for architecture students.
Conclusion
The nine tools above fall into three groups. Image generators (Midjourney, ChatGPT’s image tool, DALL-E, DeepArt and ReRender AI) help you test atmosphere and presentation quickly. ChatGPT supports writing and research. Planning tools (Finch, Autodesk Forma, Archistar and TestFit) generate or check layouts against rules, data and codes. None of them replaces the design decisions you make and defend in a crit.
Pick one image tool and one planning tool, learn them properly on a real project, and record how you used them. That record is worth more in a portfolio review than a folder of generated images. For a wider look at how offices use these AI tools in daily work, see our workflow guide.
Architecture Students Hub
This guide is part of Software, Rendering and Gear, stage 4 of our student roadmap from the first diagram to the first internship. Next stage: Portfolio, CV and Internships, or browse every free tool and guide in the Architecture Students Hub.
- best AI 3D model generators architecture
- best AI concept generators architects
- best AI design generators architecture
- best AI floor plan generators
- best AI idea generators architecture
- best AI image generators for architecture
- best AI parametric design generators
- best AI rendering generators architects
- best AI visualization generators architecture
- best generative AI tools for architects




I think AI tools for architects are interesting. They seem to help with design.
The article talks about different AI tools. They might be useful for architects, I guess.
AI in architecture is a new thing. It’s nice to see how technology is changing design.