What is Multimodal Architecture: Context Creation about?

This workshop shows how to turn natural-language rules and multimodal tools into clear, context-driven design workflows.

How is the workshop structured?

Day 1 focuses on concept creation, proto-projects, and site context; Day 2 develops narrative, atmosphere, and 3D point cloud integration.

Which software will we use?

Mainly Formas.ai (basic plan will be provided), plus Processing for interactive tests and LLMs (ChatGPT, Gemini, Grok) for concept development.

What skills will participants gain?

You’ll learn to author design rules, contextualize projects on real sites, shape atmosphere with materials and lighting, and transform sketches into 3D point clouds.

Who is this workshop for?

This workshop is designed for architects and designers eager to explore AI and multimodal tools in the early stages of architectural design.

Courses Artificial Intelligence Multimodal Architecture: Context Creation with AI

Multimodal Architecture: Context Creation with AI

Name: Multimodal Architecture: Context Creation with AI
Price: 102.00 EUR
Rating: 5 (3 reviews)

This workshop teaches multimodal design workflows to turn early concepts into context-driven architectural narratives using Formas.ai and LLMs.

multimodal_architecture_thumbnail_(2025-11-1-2)_e62f68.jpg

Overview
Content
Instructors
Reviews

Nov 1, 2, 2025

13 Lessons

Intermediate

10 Hours

Certificate — Learn

English

Unlimited Access

€120.00

€102.00

Gift Course

Introduction of the workshop:

Multimodal Architecture: Context Creation is a compact, production-oriented workshop that turns today’s fragmented multimodal stack into a clear, controllable pipeline for early-stage design. You’ll author design rules in plain language and convert them into workable constraints, build proto-projects, and place them on real sites with environmental and geometric controls.

The workshop will focus on Formas.ai (basic plan included), with Processing for interactive tests, concise ComfyUI demos to read node graphs, and LLMs (ChatGPT, Gemini, Grok) to push conceptual depth while staying grounded in structure, material, and space.

You will orchestrate prompts, masks, depth, and point clouds, then shape lighting, texture, and shot sequence to produce clear narrative fragments.

What You’ll Learn?

Author constraint-driven concepts using natural-language “rule sets” that control massing, hierarchy, material intent, and geometry.

Contextualize proposals on real sites using site cues, program, topography, scale, and circulation to move from proto-ideas to context-driven schemes.

Shape atmosphere and narrative with lighting, materials, and camera logic to communicate intent in stills and short videos.

Bridge sketched, 2D to 3D spatial data by transforming different media to point clouds for volumetric checking, spatial cues, and camera planning.

Multimodal Architecture: Context Creation with AI

Methodology:

The workshop follows a hands-on, iterative methodology that combines natural-language rule authoring with multimodal workflows across text, image, video, and 3D data. Participants move from abstract constraints to tangible proto-projects, situating them in real contexts using site and environmental cues.

By alternating between conceptual exploration in LLMs, interactive testing in Processing, and structured design development in Formas.ai, the process emphasizes rapid translation of ideas into controlled design outputs.

Program:

Day 1: Concept and Contextualization

Overview

The first day introduces the spectrum of multimodal models such as text, image, video, point clouds, and image-to-3D. Participants will receive access to the Formas.ai platform (basic plan) and explore how to use it in combination with Processing for interactive elements. Brief demonstrations of ComfyUI workflows will provide insight into advanced pipelines. We’ll also introduce large language models (ChatGPT, Gemini, Grok) as conceptual development tools.

Content

Concept Creation + Instant Interfaces

Hands-on exploration of “fluid” or “instant” interfaces.
Using natural language to define constraints and rules.
Early materialization of conceptual designs through rules and design constraints.

Turning Concepts into Proto-Projects

Developing fast, low-resolution design experiments from the constraints and rules defined earlier.
Iterating through multiple options to test feasibility and expression before full contextualization.

Beyond Images: Context Creation

Situating projects within contextual, site, and environmental constraints.
Transitioning from proto-projects to fully context-driven design proposals.

Day 2: Narrative and Atmosphere Creation

Overview

The second day shifts focus toward storytelling and atmosphere. We’ll use video-based workflows to shape design narratives, refine stylistic choices, and create immersive atmospheres. Emphasis will be on interior mood-making through light, material, and detail, while also extending into point cloud and 3D-based workflows.

Content

Narrative and Style Development

Using video models to build and control design narrative: Timelapses, Constructive Sequences, Design Rationale, Building in use.
Crafting interior atmospheres by combining light, texture, and stylistic elements. Moodboard, furniture.
Personal touch: training our own model sets and applications.

Point Cloud and 3D Integration

Converting images into point clouds for immersive spatial exploration.
Contextual creation in 3D : world generations.
Applying point clouds and image-to-3D models in design scenarios.

What You’ll Use?

ChatGPT

Formas.ai

Highlights:

Content

Introduction + PAACADEMY Updates (08:35)

From prompts to prototypes (22:31)

Evolution into AI design systems

This session introduces the evolution of computational design tools toward multi-model systems for AI-driven concept-to-3D workflows.

Concept to fabrication pipelines (59:28)

From images to spatial workflows

This session explains how current AI platforms have moved beyond simple image generation into multi-model workflows that support spatial logic, 3D geometry, point clouds, and fabrication-ready outputs.

Sketch to context generation (56:45)

Interactive AI editing workflows

This session demonstrates how to edit, crop, restyle, and re-prompt inside Freeform and Chat in FORMAS AI to rapidly iterate on sketches, shapes, context and atmosphere

From sketch to refinement (49:18)

Style-driven iterative visual refinement

This session shows how to use FORMAS AI to crop, upscale, blend, and apply style transfer to images and sketches, enabling fast iteration from early conceptual visuals into more refined architectural outcomes.

Style transfer to floor plans (48:57)

From drawings to visuals

This session shows how style-transfer, upscaling, transform prompts, background removal, and Rhino-to-Freeform vector import are used to iterate images, generate time-lapses, modify geometry

Context generation (39:28)

From sketches into atmospheric scenes

This session showcases how minimal massing models can be used to generate interior context, atmospheric lighting, and camera-based viewpoints.

Spatial imagery (57:12)

3D workflows from simple massing

This session begins to expand into 3D model and point-cloud generation, floor-plan interpretation from sketches, and the import, editing, reframing, and iterative generation of Rhino-based linework into architectural imagery.

Camera based visual generation (44:39)

Blending models into architectural scenes

This session shows how to blend model outputs, cut and collage elements, adjust point of view, switch between cameras, and generate atmospheric architectural variants directly from both 2D Freeform compositions and 3D model views.

Context driven model generations (47:02)

Direct sketch control in 3D

This session explores how to control 3D model interpretation, experiment with lighting and rendering styles, sketch directly on camera views to steer form intent, and merge generations.

From images into models (53:22)

Image-to-3D model generation

This session teaches how to generate 3D models directly from images, manage facet density, import the resulting mesh into Rhino or the built-in 3D studio, and reduce mesh complexity for smoother navigation.

Point-clouds for spatial inference (45:55)

Extending context from scans

This session goes deep into point-cloud generation from single images and videos, shows how to re-orient and rescale these point clouds in Rhino, explores extending missing context from partial captures, and compares point-cloud vs 3D model use-cases.

Controlled transitions for storytelling (43:56)

From frames into narrative video

This session demonstrates how in-between frames can be used to build controlled video narratives, how HD upscaling enhances outputs for presentations.

Screenshots:

Instructors:

Carlos Bañón

Biography

Carlos Bañón is a pioneering architect and professor integrating AI, computational thinking, and digital fabrication into architectural design. Renowned for his resource-efficient, computationally-driven methodologies, he has reshaped urban landscapes and the built environment over twenty years through his practice, Subarquitectura. As a tenured faculty member at SUTD, he established the Architectural Intelligence Research Lab (AirLab) in 2015. His acclaimed work, showcased in two editions of the Venice Biennale, garnered prestigious awards like the Singapore President’s Design Award and a Nomination for the Mies Van Der Rohe Award. An influential figure in architectural education, he held visiting professor positions at leading institutions such as MIT, EPFL, and KEA, inspiring the next generations with innovative, tech-centric designs.

Showcase

580 views

3 reviews.

MARINA LAMPRINOUDAKI 2025-11-03 17:32

It was great to hear Carlos and all the way to be more creative through the new platform he created. I found it very inspiring, and I can not wait to use it. It is defently hard work and it deserves to get a hold of all the upcoming architects.

Walter Patrick Smith 2025-11-03 18:06

Magical ahead of his time, thank you for sharing so much

Eric R 2025-11-06 22:15

Such a nice course with one of the top architects out there. The platform is very dynamic and robust, there's a lot of potential to be unlocked. While I'm satisfied with the topics and tools that were covered in a relaltively short amount of time, I still feel like the price was a little too high. Still thankful for an amazing workshop and, I just want to add, that the open forum on discord is definitely a nice follow up for the workshops!

This price is available for members only
%15.00 discount on all workshops
Access previous member-only workshops

Nov 1, 2, 2025

13 Lessons

Intermediate

10 Hours

Certificate — Learn

English

Unlimited Access

Carlos Bañón

Computational Designer, and Educator

Buy as a gift

Included in a Value-Packed Bundle!

Monthly -30%

November 2025 Workshops Bundle

Access all PAACADEMY’s November workshops in one bundle and explore multimodal design, AI ideation, reve...

View Bundle

€350.00

€245.00

🌟 Empower with Team Access

Unlimited access to the full archive
Flexible seats for evolving teams
Continuously updated course library

Contact Sales

Multimodal Architecture: Context Creation with AI

Introduction of the workshop:

What You’ll Learn?