Last month, an AI fitness app called Gravl pulled in over $440,000 in revenue. It was built by Julian Gargicevich, a developer from Argentina who started out as a guy lifting weights at the gym and ended up running a fast growing consumer app.
This is the story of how he spotted a gap in existing fitness apps, built an MVP in a few months, found his first users on Reddit, and scaled paid ads to reach over 70,000 subscribers and more than $400K per month.
If you want to build your own AI-powered app or consumer product, this breakdown will show you what actually happened behind the scenes, from idea to tech stack to ad spend and margins.
Julian's Journey: From Gym Kid in Argentina to AI Fitness Founder
Julian grew up in Argentina around sports and gyms. His dad owned a fitness center, so outside of school he was always playing sports or hanging around the gym. That early love for training stuck with him.
He later studied software engineering, then moved to Australia to work in tech. Over the years, he worked across:
- Small startups
- Large tech companies like Atlassian
- TV channels
- An investment fund
So he had two strong threads running in parallel: fitness and software.
At some point, he teamed up with his best friend and another partner to start a business. The idea was an influencer marketing platform for mobile games. Over time, it turned into a marketing agency, which none of them really wanted.
The positive side was that they learned a lot about:
- User acquisition
- How to run performance marketing
- What kind of revenue numbers successful apps were doing
After seeing those numbers up close, Julian had a simple realization: "Apps make a lot of money, especially fitness apps."
With his background in fitness and engineering, building his own app felt like the obvious next move.
If you want to follow him, you can find him on X at Julian Gargicevich on X.
Discovering the Idea: When a Fitness App Looked Good but Felt Wrong
Julian did not start with Gravl as it is today. His first instinct was to build a workout tracker similar to apps like Hevy or Strong.
He started coding a tracker, but as he built it, it did not feel right. It was basically a copy of existing apps and he could not see how it added real value. No strong hook, no unique reason to switch.
Then a friend showed him Fitbod, one of the bigger workout apps on the market. Fitbod generates workouts for you on the spot, so you do not have to plan your own routine. At first, Julian loved the idea and interface.
Once he started using it more, he noticed some issues. The workouts felt off and, in some cases, even unsafe.
In his words: "Wow, okay, these workouts are actually bad."
He saw a gap. The concept and UI were great, but the training logic did not match real, safe strength programming.
He realized there was room for:
- A similar clean and friendly UX
- Backed by a proper workout engine built on sound training logic
That was the spark. The idea became clear: build something that feels as smooth as Fitbod, but with workouts that an experienced lifter and coach would trust.
So he and his co-founders decided to go all in on an AI fitness app focused on smart, safe strength training for the gym.
You can see what that turned into here: Check out Gravl (use code STARTER).
Building the MVP: From Simple Tracker to Smart AI Coach
Julian built the first version of the app in about two to three months. The MVP had two distinct phases.
Phase 1: Simple Workout Tracker
The early version was basically a standard workout tracker:
- Log sets and reps
- Track progress over time
- Similar flow to other popular lifting apps
This was quick to build, but it did not feel unique. That is when the Fitbod insight pushed him into phase 2.
Phase 2: Intelligent Workout Engine
When he decided to mimic the good parts of Fitbod while fixing the bad training, things got much more complex.
Building custom workouts for each user meant handling a lot of business logic. The app needed to adapt to each person based on real life constraints, not just generic templates.
Some of the key inputs Gravl takes into account include:
- Equipment available
- Home gym or commercial gym
- Machines, free weights, or limited gear
- Training frequency and split
- How many days per week you train
- Preferred split, like push/pull/legs or upper/lower
- Goals and experience level
- Fat loss, muscle gain, strength, or performance
- Beginner, intermediate, or advanced lifter
- Personal details
- Age
- Weight
- Gender
- Consistency
Julian described this part as the most challenging. There were so many things to consider, and they all had to work together in a way that still felt simple for the user.
On top of that, Gravl uses AI in parts of the experience. The app generates personalized plans and workouts, and they make sure to highlight that it is an AI app, because that helps with conversions. People like the idea of a smart coach adapting to them.
By the end of those first months, they had an MVP that:
- Looked clean
- Generated personalized workouts
- Felt smarter than a basic tracker
Now they needed users.
Launching on Reddit: How a Single Post Brought the First Thousands of Users
Instead of spending money right away, Julian went where his likely early adopters were hanging out: Reddit.
He is a big supporter of Reddit as a distribution channel, especially for technical or niche products.
At that time, the app was called Gains AI, not Gravl. He wrote a detailed thread about how he built it, focusing on:
- The technical side
- The AI elements
- What the app did
Developers and tech savvy gym goers loved it.
The results from that single post:
- A few hundred upvotes within hours
- Around 300,000 impressions
- The first couple thousand users (he does not remember the exact number, but it was a lot)
Those early users were perfect for feedback. Many were developers who also liked the gym but found it intimidating or hard to program for themselves. They were willing to report bugs and suggest features.
That community feedback gave the team confidence that they were onto something. It also pushed them to treat it as a real business.
At that point, the app was free. After seeing traction and hearing from users, they decided it was time to:
- Add a subscription model
- Start running paid ads
Julian said that phase, seeing strangers use and care about his product, was one of the biggest highs of his career.
Scaling to $400K/Month: The Marketing Playbook Behind Gravl
Reddit gave them early users and validation, but it is not what took Gravl to over $4 million a year.
The real growth driver has been paid advertising.
Step 1: Turn on Ads and Watch for Signals
Once they added subscriptions, they started running ads. Julian remembers the moment they turned on their first campaign.
They got a paid subscription in the first 10 minutes.
That quick signal told them:
- The value prop was clear enough
- The pricing was not a blocker
- The funnel, even in its early form, could convert cold traffic
That early win was a huge emotional boost.
Step 2: Go Where Ads Are Cheaper
Next, they translated the app into Spanish and started running ads in South America.
Reasons this worked so well:
- Both Julian and his co-founder Matias speak Spanish
- Ad costs in South America are much lower than in the US
- There is still strong demand for high quality fitness products
They were spending less than $50 per day on ads in the beginning. That modest budget was enough to acquire users profitably in those markets.
They have continued using paid ads as their main growth channel from that point up to today.
Their main distribution channels include:
- Meta (Facebook and Instagram) ads
- TikTok ads
- Some Google and Apple Search ads
Paid traffic is not their only tool, but it is their largest driver of new users.
If you are trying to figure out where AI could fit into your own product or marketing stack, Julian mentions a resource for founders that answers that exact question: the AI Adoption Playbook from HubSpot for Startups. You can get it here: End AI Analysis Paralysis with HubSpot's AI adoption playbook.
5 Practical Tips for Building Consumer Apps in 2025
Julian shared his best lessons for anyone building apps, especially consumer or fitness apps.
Tip 1: Validate before you spend on ads
For him, validation is not just “do people like the product.”
You want to know:
- Does the product work well enough that people stick around?
- Are people willing to pay for it?
Free usage is nice, but real validation happens when customers pull out their card.
Tip 2: Use your strengths and go where competition is lower
Julian and Matias speak Spanish, so translating the app and running ads in Spanish speaking countries was an easy win.
The US is a huge market, but:
- Ads are expensive
- Competition is intense
If you have language skills or cultural knowledge, you can get better returns by focusing on markets where:
- Ad costs are lower
- Good products are still rare
Tip 3: Start small with content and creators
You do not need a massive budget to start testing creative.
Some of Julian's points:
- You can pay small creators as low as $50 for a piece of content
- You can use AI tools to generate video scripts or visuals
- You can edit videos yourself using tools like CapCut
The challenge is not making one perfect ad. The hard part is producing enough volume of creatives to test.
Tip 4: Treat UGC as your main ad style
Julian is clear on this: UGC is king.
Most of the best performing ads for consumer apps look like:
- A real person talking to the camera
- Simple, authentic explanations
- Short clips that feel like normal platform content
Now we are also seeing AI generated videos in this style. They are cheap to produce and very easy to test at scale.
Tip 5: Copy what works in your niche
You do not have to guess what type of ads might work.
With the Meta Ads Library, you can:
- Search your competitors
- See what ads they are running
- Notice which creatives stay active for a long time (usually a sign they perform well)
Julian's advice is simple: do not try to be too original at first.
Start by:
- Copying what is already working in your category
- Adapting it to your app and message
- Testing from there
There is no secret sauce. You learn by running real traffic and watching the numbers.
Inside Gravl: How the App Actually Works
To understand why Gravl converts so well, it helps to walk through the experience from a new user's point of view.
Step 1: Onboarding and Questions
Right after downloading the app, you land on a simple welcome screen. The app starts by asking for:
- Your name
- Why you want to use a fitness app
- Your training experience (for example, beginner or advanced)
If you say you are advanced, you get more specific questions, like:
- Your one rep max for key lifts
- Your goals (strength, muscle, etc.)
- How many days per week you train
You also answer questions about:
- Training frequency and preferred split
- Muscles you want to focus on
- Muscles you want to avoid for now
- Where you train, so the app can adapt to your equipment
You can set up a gym profile so Gravl knows exactly what is available there.
Step 2: AI plan generation and paywall
After onboarding, the app shows a screen where it is generating a custom workout plan.
Then you see a summary of your plan.
Right after that, you hit a hard paywall before you can access the full product. Many founders worry that users will not pay before seeing everything. According to Julian, a lot of them do.
That paywall is a key part of how the app generates so much revenue.
Step 3: Active workout experience
Once you are in the app, the core of the experience is the workout card.
This screen shows:
- The exercises you need to do today
- Sets, reps, and weights
- A clear flow that works like a workout tracker
Each exercise includes:
- A video
- A text description
- Tips on how to perform it correctly
There are also smart features, such as:
- If you reorder exercises, the app adjusts the weights based on the new sequence
- A recovery rate that estimates how tired your muscles are, using both your Gravl workouts and data from Apple Health or Strava
So if you went for a long run or ride using another app, Gravl can adjust your next strength workout based on that.
Human support built in
Inside the app, they also offer 24/7 support with:
- Help articles
- Direct messages with an actual person
Julian stressed that support is not handled by AI. Users value knowing that a real human will reply when they have questions or issues.
All these pieces together explain how Gravl keeps users engaged long enough to support $400K+ in monthly revenue.
Tech Stack and What It Costs to Run Gravl
Julian also shared what Gravl is built on and how their costs break down.
Tech stack
On the development side, they use:
- React Native and Expo for the mobile app
- .NET for most of the core backend functionality
- Next.js and React for internal admin dashboards
For AI assisted coding, Julian uses tools like Cursor, but he is very careful about how he uses them. He directs the AI to specific files and avoids letting it change too much at once.
Main expenses
Julian recorded a follow up segment to share financial details.
Here is how their expenses roughly break down:
| Category | Notes |
|---|---|
| Paid ads | About one third of revenue, mostly Meta and TikTok, plus some Google and Apple Search |
| Salaries | Started as 3 founders, now a team of about 13 to 14 including part timers and contractors |
| App store fees | Around 15% cut from Apple on in-app purchases |
| Attribution / MMP tools | Around $1,000 per month |
| Infrastructure and AI tools | Around $1,000 per month for servers, AI bills, and other tools |
He mentioned salaries in the range of $50K to $80K, though context suggests that covers either individual ranges or perhaps a rough band for some roles, not the full payroll number.
Even with these costs, spending only about one third on ads plus operational expenses leaves healthy margins at their current revenue level.
Also Read: From Garage Dreamer to Billion-Dollar Visionary: The Jamie Siminoff Story You Haven’t HeardFinal Advice from a $400K/Month Founder
Julian closed with a few simple but strong points for builders.
First, be proud of what you build. For him, working on something he is excited about, like an AI fitness app tied to his love of training, cannot be matched by anything else. It makes it easier to grind when the money is not there yet.
Second, keep pushing. Building the product is not even half the work. You still have to:
- Post on Reddit
- Run ads that fail
- Fix bugs for real users
You will take hits. You just have to keep going.
At the same time, you need to know when to quit. Their previous startup, the influencer platform turned agency, dragged on for about a year longer than it should have. Looking back, he believes they should have shut it down sooner and moved on.
The reason Gravl works so well is that it sits at the intersection of what Julian is good at and what he loves. It helps hundreds of thousands of people work out smarter, and it pays the bills in a serious way.
If you want to build something similar with AI, you can check out the program created by Starter Story that helps founders ship fast: Build your own app with AI through Starter Story Build.
Building an app people pay for is not easy, but Julian's story shows that focus, patience, and real user feedback can turn a gym idea into a $400K per month product.
0 Comments