How I Built a Real Flutter Mobile App With Primio - Here’s What Happened
A while ago, I tried Darvin, an AI app builder that could turn an idea into a mobile app. I also wrote about my experience with it on this blog.
Darvin has changed a lot since then. It is now called Primio, and I wanted to see how much the product had actually improved.
Instead of going through its website and writing another list of features, I decided to build something with it.
I also wanted the project to be related to something I actually deal with as a content creator. That is how I ended up building CreatorDeal, a small app for tracking brand collaborations.
Here is how it went.
What Is Primio?
Primio is an AI app builder that generates Flutter projects from prompts.
The basic flow looks like this:
Prompt → Flutter project → Preview → Iterate → Build for mobile or web
You describe what you want to build, Primio generates the project, and then you can keep working on the same app through chat.
There is a live preview for quickly checking changes, and the project can later be built for Android, iOS, or the web.
That sounded interesting, but I was more interested in seeing what the process was like in practice.
Want to try building your own app? You can explore Primio and start a project at Primio.dev.
Why I Built CreatorDeal
Brand collaborations can get messy quite quickly.
One conversation starts through email, another one happens on Discord, and sometimes a brand reaches out through a DM. Then I still have to keep track of things like deadlines, deliverables, payment status, notes, and what stage each collaboration is currently in.
I wanted a simple place where all of that could live together.
That became the idea behind CreatorDeal.
It is basically a sponsorship tracker for creators, built around this workflow:
Inquiry → Negotiating → Confirmed → Creating → Published → Paid
I deliberately did not make a to-do list, habit tracker, or notes app. Those demos are everywhere, and I wanted to give Primio something a little more specific.
CreatorDeal would keep track of brand names, deliverables, fees, deadlines, collaboration status, and the next thing that needs to be done.
With that idea in mind, I started a new Primio project and gave it my first prompt.
Starting With the First CreatorDeal Build
I left the backend out on purpose for this first version.
Before thinking about authentication or databases, I wanted to see how Primio handled the basic product idea, navigation, and UI.
This was the prompt I used:
Create a mobile-first application called CreatorDeal.
CreatorDeal is a brand collaboration and sponsorship tracker for content creators, YouTubers, bloggers, influencers, and freelancers.
Its purpose is to help creators manage brand partnerships from the first inquiry until the content is published and payment is received.
CORE WORKFLOW
Use these collaboration statuses:
- Inquiry
- Negotiating
- Confirmed
- Creating
- Published
- Paid
CORE EXPERIENCES
Build:
1. Home
2. Deals
3. Deal Details
4. Add/Edit Collaboration
5. Profile
Use bottom navigation only for:
- Home
- Deals
- Add
Profile should be accessible from the user's avatar in the top-right corner.
DESIGN DIRECTION
Create a premium, editorial, creator-first mobile product inspired by the visual philosophy of modern creator partnership platforms such as Kolect.
Do not copy any specific Kolect layout, logo, proprietary component, branding, text, or visual asset.
The inspiration should only influence the overall product direction:
- creator-focused
- clean
- spacious
- earnings-oriented
- premium
- editorial rather than dashboard-heavy
Avoid the appearance of:
- a generic admin dashboard
- an enterprise CRM
- a generic AI-generated SaaS interface
- excessive cards
- excessive gradients
- heavy shadows
- colorful analytics dashboards
VISUAL SYSTEM
Use:
- warm off-white main background
- white surfaces
- near-black typography
- generous whitespace
- subtle dividers and borders
- restrained rounded corners
- almost no heavy shadows
- oversized financial figures
- typography-driven hierarchy
- one acid-lime accent color used sparingly
- muted pastel colors for statuses
- editorial section headings
- concise creator-focused microcopy
HOME
At the top show:
CreatorDeal wordmark
User avatar
Greeting:
"Good morning, Abdul"
Supporting copy:
"Here's what's moving."
Create one prominent dark near-black earnings card.
Show:
THIS MONTH
$2,450
potential earnings
$620 awaiting payment
Below the earnings card show compact indicators for:
7 Active Deals
2 Due Soon
Create a "Deals in progress" section using compact editorial rows.
Each collaboration should show:
- brand monogram
- brand name
- deliverable
- fee
- status
- deadline
Create an "Up next" section showing upcoming actions and deadlines.
DEALS
Show:
- number of collaborations
- total pipeline value
- simple filtering
DEAL DETAILS
Show:
- brand
- deliverable
- fee
- status
- deadline
- contact details
- notes
Add a vertical progress timeline:
1. Brief received
2. Contract confirmed
3. Product received
4. Content created
5. Draft submitted
6. Client approved
7. Published
8. Payment received
ADD COLLABORATION
Use a four-step flow:
Step 1 — Brand
Step 2 — Deliverable
Step 3 — Deal
Step 4 — Review
Use realistic fictional sample data.
Do not add:
- authentication
- Supabase
- external APIs
- payments
- notifications
- AI assistants
Use local sample data only.
My first version of CreatorDeal generated from the initial prompt.
The first result surprised me.
Primio did not just give me a couple of disconnected screens. The main structure of the app was already there, and it looked much closer to a real mobile product than I expected from the first generation.
It was not finished, of course, but there was already enough there that I could start thinking about improvements instead of starting over.
Asking Primio What Was Wrong With Its Own Work
One feature I found useful during this process was Discuss mode.
Instead of making changes as soon as I send a message, Discuss mode lets me talk about the project first.
I decided to use it in a slightly unusual way: I asked Primio to review the app it had just generated and tell me what could still be improved.
This is what I asked:
Do not modify the project.
Review the current CreatorDeal implementation from a mobile product design and UX perspective.
Evaluate how closely the current implementation matches the intended direction:
- premium creator-first product
- editorial rather than dashboard-like
- generous whitespace
- typography-driven hierarchy
- restrained use of cards
- restrained acid-lime accent
- prominent earnings information
- compact collaboration rows
- clear next actions
- visually distinct from a generic SaaS admin dashboard
Identify:
1. What currently works well
2. What still looks generic
3. Where visual hierarchy is weak
4. Where there are too many cards or containers
5. Typography problems
6. Spacing problems
7. Mobile usability issues
8. Any unnecessary UI elements
9. What should be improved next
Do not write or modify any code.
Only analyze the current implementation.I liked being able to do this before changing anything.
It gave me a chance to think about the next step without risking a redesign of parts that were already working.
Once I had gone through the feedback, I switched back to editing and focused on the interface.
Refining the CreatorDeal Design
I did not want a completely new design. The first version already had the right structure, so the next prompt was mostly about cleaning things up.
I wanted better spacing, stronger hierarchy, fewer unnecessary cards, and a clearer focus on earnings and active deals.
Now refine the existing CreatorDeal interface.
Do not change:
- information architecture
- navigation
- existing data
- current screens
- current functionality
Focus exclusively on visual hierarchy, layout, typography, spacing and interaction design.
HOME
Make the earnings area the strongest visual anchor.
Use one near-black card containing:
- large potential earnings figure
- small "potential earnings" label
- awaiting payment underneath
Use acid lime only as a restrained accent.
Do not place every metric inside an individual card.
Show:
Active Deals
Due Soon
as compact text-based metrics below the earnings card.
Make "Deals in progress" an editorial list rather than a card grid.
Each deal row should prioritize:
Brand name Fee
Deliverable
Status · Deadline
Use subtle dividers between deals.
DEALS
Make the collaboration list more elegant and slightly denser.
Brand and fee should be visually dominant.
Deliverable, status and deadline should remain secondary.
Avoid unnecessary shadows and containers.
DEAL DETAILS
Use this visual hierarchy:
Brand
Deliverable
Fee
Status
Next Action
Progress
Details
Make the progress timeline vertical and visually lightweight.
GENERAL
Improve:
- typography hierarchy
- vertical rhythm
- horizontal alignment
- whitespace
- touch targets
- consistency
Preserve:
- warm off-white backgrounds
- near-black typography
- white surfaces
- subtle borders
- restrained acid-lime accents
Do not add new functionality.
The CreatorDeal home screen after another round of design refinement.
The useful part here was that I did not have to explain CreatorDeal from the beginning again.
I could point Primio at one part of the existing app, ask for a change, check the result, and move on.
Adding Search and Deal Filters
Once the UI was close to what I had in mind, I moved on to something more functional.
The Deals page already showed the collaborations, but I wanted it to be easier to use when the list started getting longer.
Search and filtering felt like a good next test.
This was the prompt:
Improve the existing Deals experience.
Do not redesign unrelated screens.
Keep the six underlying sponsorship statuses:
- Inquiry
- Negotiating
- Confirmed
- Creating
- Published
- Paid
At the top show:
Deals
[number] collaborations
[pipeline value] in pipeline
Add four primary filters:
- All
- Active
- Negotiating
- Completed
Use these definitions:
All:
all collaborations
Active:
Confirmed or Creating
Negotiating:
Negotiating
Completed:
Published or Paid
Add a secondary filtering option where the user can select any individual detailed status.
Make filtering functional using the existing local data.
Add search.
Search should match:
- brand name
- deliverable
Keep the current editorial product design.
Do not introduce:
- data tables
- CRM-style dashboards
- large filter panels
- additional navigationThe search worked with the sample data, and the filters also changed the list based on the selected category.
That was the point where CreatorDeal started to feel less like a static demo and more like something I could actually interact with.
If you want to experiment with the same kind of prompt-based app building, you can create your own project at Primio.dev.
Testing CRUD With Local Data
Search and filters were useful, but I wanted to go one step further.
Could I add a new brand collaboration, open it again, change its details, and delete it?
That meant testing the basic Create, Read, Update, and Delete flow.
Here was the prompt:
Make the existing Add Collaboration workflow fully functional using local application state.
Do not add any backend yet.
Preserve the current UI.
STEP 1 — BRAND
Required:
Brand name
Optional:
Contact name
Contact email
STEP 2 — DELIVERABLE
Options:
Dedicated YouTube Video
YouTube Integration
Instagram Reel
TikTok Video
Blog Article
Newsletter Mention
Other
STEP 3 — DEAL
Fields:
Fee
Currency
Deadline
Status
Notes
STEP 4 — REVIEW
Show:
Brand
Deliverable
Fee
Currency
Deadline
Status
Provide:
Back
Edit
Add Collaboration
AFTER SAVE
The newly-created collaboration should:
- appear in Deals
- open correctly in Deal Details
- update pipeline value
- update active deal count when appropriate
- update earnings metrics when appropriate
Make these operations functional:
Create
Read
Edit
Delete
Keep everything stored locally for now.
Do not add authentication or Supabase.Testing the CreatorDeal workflow using locally stored data.
This was the point where I could add and change actual deal data instead of just clicking through generated screens.
Adding a collaboration could also update other parts of CreatorDeal, including the deal list and the summary numbers on the home screen.
I stopped here for this first experiment. Everything in this version is still stored locally.
What I Haven't Tested Yet: Supabase
That local storage is also the biggest limitation of the CreatorDeal version shown here.
I would not call this a production-ready app yet, because there is no real backend or user account system behind it.
Primio also supports backend workflows with services such as Supabase. That would make it possible to add things like:
- User accounts
- Email and password login
- Persistent deal data
- A real cloud database
- Separate data for each user
- Progress that remains available after closing the app
That is probably the next step I would take with CreatorDeal.
For this article, though, I wanted to keep the test focused. I was mainly trying to answer one question:
Can I take a creator-focused app idea, describe it in normal language, improve the UI, and add working interactions without writing the Flutter code myself?
For this version of CreatorDeal, I was able to get far enough to answer that question without adding a backend yet.
Running CreatorDeal in the Primio Emulator
The last thing I tried was the Primio Emulator.
The normal Live Preview had been enough while I was changing the UI, but I also wanted to see CreatorDeal running inside a device environment.
So I created a build and opened it in Primio's Emulator.
CreatorDeal running as a built application inside the Primio Emulator.
For this project, I mainly used the Emulator to check how the app looked and behaved outside the usual preview.
Primio also exposes device logs, which should be useful if a future Android or iOS build behaves differently from the preview.
So, What Do I Think About Primio?
After building CreatorDeal, I think Primio works best when I use it step by step.
The first prompt gave me the starting point. After that, I could review what it created, discuss the design, improve a specific screen, and then add functionality without rebuilding the whole project.
That workflow made more sense to me than expecting one huge prompt to produce a finished app.
CreatorDeal is still a small project. It does not have a real backend yet, and I would not call this version a finished product.
Still, it went beyond a static mockup. I ended up with a mobile app concept where I could search deals, filter them, add new collaborations, edit data, and test the result inside an emulator.
The part I want to try next is Supabase. If I can add authentication and persistent data without having to rebuild CreatorDeal from scratch, that will be a much more interesting test of how far Primio can go.
Want to Try Primio?
If you already have an app idea and want to see how far you can get without starting a Flutter project from scratch, you can try the same process yourself.
Start with the first version, see what Primio gives you, and then improve it one part at a time instead of trying to solve everything in a single prompt.
