FreewayML | Sebastian Gamarra

Designing Freeway’s Brand: A visual identity for endless creativity

When I joined FreewayML, our initial focus was on designing the app’s first wireframes. But soon after, we shifted gears to define the brand’s identity. The founders had a clear vision: they wanted the brand to embody the concept of freeways, open roads with endless possibilities.
‍
Why a freeway? Much like a freeway offers multiple routes to a destination, FreewayML allows users to explore countless paths to creativity. To capture this idea, I designed the brand around an aerial view of freeways, blending it with the infinity symbol to reinforce the concept of limitless imagination. The choice of deep purple as the primary color further emphasized a sense of mystery and creative magic.

Freeway idea exploration around actual freeways from above and combined with the infinity sign.

Freeway final logo variants.

What

In the early days of AI image generation (Stable Diffusion v1, DALL·E 2), most available tools were either overly complex or lacked essential features. Many had cluttered UIs, required advanced knowledge of prompt engineering (e.g., ar 9:16, f2.0), and offered no cloud storage for generated images. This made the creative process inaccessible to many users.

How

We took a mobile-first approach to design a streamlined, user-friendly interface with minimal frontend controls while leveraging advanced backend processes. By integrating expert prompt optimization, model-chaining, and performance enhancements, we ensured that users could generate high-quality images effortlessly, without needing deep technical expertise.

Why

FreewayML was built on the belief that AI should enhance creativity, not replace it. At the time, there was no middle ground between full professional control and a simplified user experience. Our goal was to bridge this gap, empowering users with powerful AI tools while keeping the process intuitive and accessible.

My design flow focuses on rapid prototyping and iterative improvements, ensuring the best possible results through continuous refinement.

Crafting the Core of FreewayML: The ‘Studio’ Tab & Image Editor

Achieving simplicity in design is often the most challenging task, especially when a system offers endless possibilities. As a designer, the key is to establish a clear hierarchy that aligns with user needs while keeping the experience intuitive.
‍
This was the core challenge in designing FreewayML’s UI. We began by benchmarking existing solutions, not just within the AI space but across various industries, to identify and integrate the best design patterns. From there, the foundation took shape: a streamlined, full-width prompt bar that made image generation effortless.

FreewayML benchmarking, identifying the pros and cons of industry leaders.

With this direction in place, we had a clear path to designing FreewayML’s core feature, the image generator. We started with the essentials: a simple way to enter a prompt and generate images, along with basic editing capabilities. Below are the first iterations of the “Studio” tab, which served as a foundation for learning, refining, and continuously improving the experience.

First iterations of the Studio tab, the core feature of FreewayML

First iterations of the image detail view, and the image editor.

As AI image technology evolved rapidly, we introduced Quick Actions, a set of intuitive controls that allowed users to perform instant transformations like Enhance Image, Create Variations, and Image-to-Image. Following this, we planned to integrate AI Layers (or Masks) to expand creative possibilities.

Final iteration of the Image Editor, merging the Viewer and Editor into a seamless experience. With a mobile-first approach, we ensured better scalability, making it easier to integrate new features as the app evolved.

Final desktop iteration of the Image Editor

Refining the Studio Tab: Streamlining the Creative Flow

With the core functionality in place, it was time to refine the “Studio” tab to ensure a smoother user experience. While we were close to our goal, allowing users to effortlessly create and edit images, there were still a few pain points that needed to be addressed:

Lack of visible history – Users had to switch to the Assets tab to compare previous generations or revert to an earlier version, disrupting creative flow.
Unclear past actions – There was no indication of whether a previous generation was created using Variants or Image-to-Image.
Limited image previews – The system only supported up to four images, restricting creative exploration.
Confusing hierarchy – The Original image was displayed with the same prominence as new results, causing user confusion.
No prompt tracking – Users had no way to view or copy past prompts they had used.

By analyzing user sessions with LogRocket, we identified these friction points and gathered valuable insights. With this data in hand, it was time to iterate and refine the experience further.

Final iteration of the Studio tab, featuring image history, action badges for better clarity, and several usability enhancements.

Improvements to the Style drawer.

Walkthrough about using masks.

Screens showcasing the public Community profile and user posts.

Main screen and detail view of Freeway Challenges.

Assets main view and folders detail view.

Upload an image for Face animations.

Showcasing the AI model training process, Image selection, training initiation, and the public AI model detail page.

Sketching process over a mask.

Marketing site landing page designed in Figma then developed using Webflow

Landing page highlighting the ‘Blend’ feature and its capabilities. Developed using Webflow.

But did all this hard work pay off? Let’s take a look at the results. 📈

FreewayML experienced incredible growth in just a few months, from its launch in September 2022 to November 2022, where we saw record-breaking signups, daily operations (such as Blend, Enhance, Variations, and more), and various other performance metrics.
‍
What’s even more impressive? This was achieved with zero ad spend—just organic sharing across platforms like Reddit, X, and more.

As shown in the graphs, we saw a significant surge in activity whenever we rolled out new updates. Additionally, the FreewayML Discord community grew to over more than 10,000 active users, where they could generate images using the FreewayML Bot. This not only created an engaged user base but also provided us with instant feedback, allowing us to continuously iterate and improve the app.

Even MattVidPro reviewed the app!

MattVidPro a popular AI and technology YouTuber with over 280,000 subscribers, took the time to review FreewayML. In his video, he shared his honest thoughts on the app and its features. Be sure to check it out!

"I would say, so far this interface is already putting ahead of the pack, easy to use interface, really easy to understand. Snappy, clean UI."

MattVidPro AI

Unleash Your Artistic Side with FreewayML's AI Editor @ 2:51

A shared vision driving the success of our partnership

The comprehensive design overhaul and brand elevation of FreewayML have successfully repositioned the platform as a leader in the AI image space, transforming its visual identity and significantly boosting user engagement.

Working alongside a small, highly synchronized team enabled us to achieve remarkable results. Revolutionizing the AI image industry and generating a lasting impact was an exciting, first-time experience for me, and one that I truly enjoyed.

Let’s collaborate and create something amazing!

Reach out to explore business opportunities or simply say hello! Use the form below or copy my email to get in touch.

stgamarra5@gmail.com

Download CV

Thank you! Your submission has been received.
I’ll be in touch with you soon ✌️

Oops! Something went wrong while submitting the form.