Built a chat-based AI expense tracker. Coding was easy—now how do I scale traffic?
Like many engineers, coding was the easy part.
Over the last few months, I built and launched Vitmora — a privacy-first, chat-based AI expense tracker built for a global audience.
The Core Idea
Most finance apps ask users to link bank accounts or allow SMS scanning. A lot of people are uncomfortable with that.
So I built a zero-friction, private alternative that works globally with support for 150+ currencies.
How It Works
-
Just type or speak naturally:
“Spent $12 on coffee”
“Paid ₹450 for groceries”
“Uber ride for €18” -
AI automatically categorizes the expense
-
Dashboards update instantly
-
No bank linking
-
No SMS scraping
-
No complicated forms
The app is now live on iOS + Android:
https://vitmora.com
Now comes the hard part: getting users
I’m running this solo with basically a $0 marketing budget and trying to figure out how to build long-term organic traffic.
For founders who’ve grown utility/mobile apps:
-
What free distribution channels worked best for you?
-
How do you create enough “digital noise” for Google and AI search engines to start recognizing a new product?
-
Any brutal feedback on the landing page or positioning?
I’d genuinely appreciate advice from people who’ve been through this phase.
Traffic's kind of the wrong thing to optimize first with a chat app. If it feels dead on first open, no amount of traffic saves you. I still have 0 onlie on my screen for way too long and it was quietly tanking conversions. Switched it to a cumulative "X chats started" and it stopped bleeding. For acquisition, Apple Search Ads Basic is boring but it works
Mobile personal finance is one of the most saturated and CAC-expensive categories in consumer software. Three structural problems no marketing strategy fixes.
Category is brutally saturated — YNAB, Copilot, Monarch, Rocket Money, Cleo, Plum, Emma, plus free Notion templates. "AI + chat-based" isn't a wedge — Cleo's done AI personal finance since 2019.
"Privacy-first, no bank linking" is weaker than it looks. Manual-entry tracking dies because people stop entering — the oldest problem in finance software. Privacy-first means more friction by design. The market that wants this already gave up on tracking entirely.
Mobile finance apps grow from ASO (incumbents own keywords), paid acquisition ($50-150 per install industry standard), or influencer partnerships. Not SEO, not IH posts. $0 against $50-150 CAC is a category mismatch, not a marketing problem.
"Global, 150+ currencies" also works against you. Successful indie finance apps start hyper-local. Global with $0 budget = $0 effort everywhere.
Real question: is this category winnable solo, or should you go hyper-narrow on one country + user type.
Thanks for taking the time to write this. You made some really valid points, especially around how crowded the category is and how hard it is to get people to keep tracking expenses long term.
The manual-entry problem is definitely the biggest challenge. My thinking was that if logging expenses feels more like sending a message or speaking naturally, maybe more people would stick with it compared to traditional forms.
I also agree that going global from day one has its downsides. Still trying to learn where the strongest demand is before narrowing the focus.
And on the privacy side, I noticed a lot of people are hesitant to connect bank accounts or give SMS access, which is one of the reasons I wanted to try a different approach.
Still very early and figuring things out, so I genuinely appreciate the honest feedback.
On the "chat makes people stick" hypothesis — worth pressure-testing because that's the core bet now.
The manual-entry problem isn't friction per entry. It's that the user forgets to open the app at all. You can make entry frictionless and it still dies when they forget they were supposed to log lunch yesterday. Apps that tried easier entry (voice, widgets, NFC tags) extended use by 2-4 weeks, then users forgot. The retention curve barely moves.
Auto-import (bank linking, SMS) wins long-term not because of UX — because the user doesn't have to remember. Privacy-first means manual means forgetting.
Worth testing the assumption now rather than in six months.