---
title: "Pedro Martins Architect Case Study | Gendo"
description: "How a Porto-based solo architect replaced a tangled ComfyUI pipeline with Gendo's canvas—freeing his hardware, opening collaboration to clients, and produ…"
lang: en
json-ld:
---

[![Gendo](https://assets.gendo.ai/marketing/homepage/gendo-logo-white.png)](/)

[Features](/product)[Solutions](/workflows)[Resources](/resources)[Pricing](/pricing)

About

[Log In](https://app.gendo.ai/sign-in)[Start creating ](https://app.gendo.ai/sign-up)

[Back to Resources](/resources)

Story 

# Pedro Martins: A Canvas-First Workflow for a Solo Practice

How a Porto-based solo architect replaced a tangled ComfyUI pipeline with Gendo's canvas—freeing his hardware, opening collaboration to clients, and producing presentation-quality visuals from real photos and 3D views.

Unlimited

Image generation

Freed-up

Local hardware

Clients in

The canvas

![Pedro Martins Architect](https://assets.gendo.ai/marketing/pedro-martins-logo.png)

About the practice 

## Pedro Martins Architect

Team Size

Solo practice

Location

Porto, Portugal

Principal

Pedro Martins

Focus

Apartment buildings, Public projects

Pedro Martins is a solo architect based in Porto, Portugal. After twelve years co-running a multi-generational family practice—where he covered every stage from concept to site execution, including the studio's 3D visualisation work—Pedro now runs his own practice independently.

His work focuses on apartment buildings designed from the ground up, alongside collaborations on public buildings. He is also one of a growing number of practitioners using AI as a core production tool, not a novelty.

> "I had to cover the rendering side because I was the youngest in the studio. That experience prepared me to be solo."
> 
> — Pedro Martins

Live canvas 

## Explore Pedro's project canvas

A live look inside the Gendo canvas Pedro shared from his own practice.

[Open in New Tab](https://app.gendo.ai/project/d258fe51-9a3a-4f2c-b9a3-c21044600964)

The challenge 

## A messy AI pipeline and a slow traditional one

Pedro's challenge was twofold. Relying on traditional 3D modelling and V-Ray for early-stage testing was too slow—it delayed internal decisions and risked clients losing confidence in the direction.

On the AI side, ComfyUI's node-based interface obscured the design process. With so many images being generated per project, it became hard to see how decisions had evolved. Generation also ran on his local hardware, blocking the same machine he needed for 3D modelling and BIM documentation.

The underlying problem: producing high-quality, fast visuals as a solo architect, without losing track of the design process or tying up the hardware he needed elsewhere.

The solution 

## A canvas that shows the process, not just the output

Gendo's canvas immediately answered the part of ComfyUI that had been failing Pedro: a clear, image-led overview of the entire decision process, with everything else tucked into the background. Combined with fine-tuned style models, intuitive annotation, and server-side image generation, it slotted directly into his solo workflow.

"Gendo seemed much easier to oversee the process of the content generated—essential for project development when you have so many images to generate."

Server-side generation also freed Pedro's local hardware for 3D modelling and BIM documentation.

Workflow 

## From photographs and SketchUp to a single canvas

Pedro's typical workflow starts with two very different inputs. For renovation projects he photographs existing built apartments and manipulates them directly inside Gendo. For new-build apartment buildings he exports views from his 3D software, takes a screenshot, and pastes them straight onto the canvas.

From there, everything happens in one place—iteration, annotation with the pencil tool, fine-tuned models for specific styles, and presentation of the results to the client. There is no shuttling between Photoshop, render queues, and a separate presentation deck.

> "I just print-screen and paste. It's so perfect for the way I work."

Before Gendo 

## Two years on ComfyUI, and the limits of traditional rendering

Pedro had been running ComfyUI for two years before adopting Gendo. While powerful, the interface became unmanageable for project work—too many nodes, too much going on, and the history of decisions lost in the noise.

On the traditional side, the speed at which you can see a design concept depends entirely on getting a 3D render. Until then, everything has to be modelled and textured, and you wait a long time before any visualisation of the idea exists. For a solo practice that has to constantly check direction with the client, that delay is expensive.

## The Impact: Faster iteration, freed hardware, and clients inside the process

Gendo gave Pedro two distinct gains: efficiency from offloading image generation to the server, and a more transparent way to develop projects with clients. Instead of producing a PDF and presenting at fixed milestones, the canvas itself becomes the live workspace.

Impact 

## Key Outcomes

Hardware freed for production work 

Image generation runs on Gendo's servers, leaving the local machine for 3D and BIM

Clients inside the canvas 

Clients can enter the canvas and explore images directly, with no PDF assembly in between

Quality clients now expect 

Replaces slower V-Ray workflows that no longer hit the time/quality bar

Generative serendipity 

Unexpected results from generation occasionally spark new directions in the design itself

Feedback 

## Transparent design development, shared responsibility

Pedro deliberately invites clients into the development of the design, the delivery timeline, and proposed changes—rather than only presenting a polished end render. The canvas makes that openness practical.

Clients are finding it intuitive to enter and browse on their own, removing the need for Pedro to assemble a PDF for every development milestone. Image quality has been consistently strong, and bringing the client into the workspace shifts decision-making into a more shared mode.

"Having a canvas where the client can participate in the process and build on what I've proposed is amazing."

> "I'd recommend stopping all standard rendering tasks. Dedicate your time fully to 3D modelling and BIM documentation, and use Gendo to generate and manage all project visualisations."

![Pedro Martins](https://assets.gendo.ai/marketing/pedro-martins-headshot.png)

Pedro Martins

Architect, Porto

## More Stories

[View all](/resources/stories)

[

![ro01.studio](https://assets.gendo.ai/marketing/ro01-forest-cabin.jpg)

### Output That Keeps Up With the Ideas

Read story



](/resources/stories/ro01-studio)

[

![Lanza Architects](https://assets.gendo.ai/marketing/lanza-hero-populated.png)

### Faster Visuals, Stronger Proposals: Lanza Architects' Gendo Workflow

Read story



](/resources/stories/lanza-architects)

[Gendo ](#)

A living canvas for visual exploration, iteration, and decision-making.

### Product

-   [Product](/product)
-   [Pricing](/pricing)
-   [FAQs](/faqs)

### Company

-   [About Us](/about-us)
-   [Career](/career)
-   [Environmental Statement](/about/environmental-statement)
-   [Blog](/resources/insights)
-   [Contact Us](/book-call)

### Connect

-   [LinkedIn](https://linkedin.com/company/gendo-ai)
-   [Instagram](https://instagram.com/gendo.ai)
-   [YouTube](https://youtube.com/@Gendo-ai)

© 2026 Gendo. All rights reserved.

[Licenses](/licenses)[Terms of Service](/terms-of-service)[Privacy Policy](/privacy-policy)