Stop Building Fintech with Databases: Why I Went Local-First for Pocket Portfolio

#fintech #privacy #localfirst #nextjs

Stop building fintech with databases. Why I went local-first for Pocket Portfolio.

For the last decade, building a fintech app meant one architecture: a centralized PostgreSQL database, a backend in Node or Python, and an API that hoards user data.

As an engineer, I always found this disturbing. Why does a portfolio tracker need my trade history stored on AWS us-east-1? It does not. It stores it because the data is the product, not the software.

I wanted to build a tool for modern investors — those of us using Trade Republic, Trading 212, or Robinhood — who want insights without surveillance.

So I set a hard constraint:

Zero user data leaves the client.

That constraint became the foundation of Pocket Portfolio — a privacy-absolute, local-first financial application.

The Zero-Server Stack

We stripped everything back.

There is no backend database.

No user accounts.

No cloud sync.

The Core: Next.js (static export). The entire app is JavaScript shipped to the browser.
The Database: Browser storage (localStorage and IndexedDB). Your financial data lives on your device, not ours.
The Parser: Web Workers processing CSV files entirely client-side.

No servers storing user data.

No silent replication.

No surveillance by default.

The Hardest Part: Client-Side Parsing

The hardest technical problem was not charts or performance.

It was standardizing messy, inconsistent CSV exports from more than 15 brokers — without sending a single byte to a server.

Each broker formats exports differently. Headers change. Dates vary by locale. Fees are embedded or omitted. Some files are technically valid CSVs but semantically broken.

So we built a robust, open-source parsing engine that runs entirely in the browser.

The flow looks like this:

User drops a CSV file (for example, a 5MB Trade Republic export).
The main thread hands the file to a Web Worker to avoid UI freezes.
The worker identifies the broker schema and normalizes each row into a standard JSON format.
Clean data is returned to the main thread and hydrated into application state.

All of this happens locally.

No uploads.

No APIs.

The Market Has Spoken

I was not sure if other developers cared about this level of privacy. I was wrong.

Within the first seven days of soft-launching the parser engine on npm, it crossed 5,800+ weekly downloads.

Developers are already integrating the local parser into their own pipelines.

The takeaway is simple:

You do not need to hoard user data to build valuable software.

Try It (and Break It)

Pocket Portfolio is live.

If you are an investor or a developer tired of the current fintech status quo, try it with your messiest CSV file and see how the local engine handles it.