This is literally how every banking API works. There are not many options aside from screenscraping.

This is fascinating. I’d always assumed that these open banking platforms had to go through a protracted period of relationship-building and negotiation with financial institutions, but you’re saying the most common approach is scraping? Is this how, e.g., Plaid does it?

Correct. And not just for banking either. For financial brokerage APIs, a lot work the same way.

That is ... absolutely bonkers.

Thanks for your insight.

Isn't this how Mint did/does it?

It's similar in a way, yes. We have clients that are just like Mint for the Latam market.

We always try to do our dev work and connect to both the website and the app. In many (most) instances, though, in Latam going via web is the only way forward. That being said, we actively work with banks to partner up and establish direct connections too.

wouldn't this be slow? Is it legal in LATAM for screen scrapping API?

Partnering with banks you mean? Yeah, banks are slow to move but in Latam they're also pretty keen to partner up with fintechs if they see a win/win.

We work on the premise of end-users granting us permission to access and share the data they want for the means/apps they want. So there're an opt-in here. There's no general legal framework against this.