Wesabe

What it is

Wesabe was a personal finance website that ran from November 2006 to July 2010. It was established in December 2005, analyzed your bank and card transactions, and gave advice on saving money, much of it from community groups where members swapped tips. Its CEO was Marc Hedlund. It is gone: the shutdown was announced on 30 June 2010 and the accounts service ended on 31 July 2010. Today wesabe.com redirects to an unrelated company (checked 10 October 2026).

It matters to ownpurse because of how it lost, and because of what its founder wrote about it.

How it works

Wesabe did not license a commercial aggregator. It built its own uploader. The open-sourced version, SSU, the Server-Side Uploader, is a scripted Mozilla browser (XULRunner) with one script per bank, filed by reverse-DNS name, that logs in with your credentials and downloads OFX. It runs a small HTTP server on your machine and answers in JSON. The idea was that bank passwords need not leave your computer, and that a real browser breaks less often than a bare scraper.

On the site you tagged transactions, edited descriptions, set spending targets and watched trends. The transaction and analytics engine (brcm-accounts-api) and the Rails front end (pfc) sat behind it.

Data and access

The record lived on Wesabe’s servers. When it closed, users were told to download their account information by 31 July, after which it would be deleted (TechCrunch). The company then released the site’s code under Apache-2.0, so a person could run it on their own Mac or Linux machine. The repositories, including fixofx, a converter from several financial formats to OFX 2, were archived on 8 May 2018. Whether Wesabe offered a public developer API while it ran is not confirmed here.

Agents

None. Wesabe predates agents and used no models. Its relevance is the uploader: per-source scripts are exactly the kind of connector an agent can now write in an afternoon.

Customers

Consumers. No user counts were published. TechCrunch reported that traffic fell steadily after its April 2009 iPhone app. Fans liked the downloadable import client and the privacy stance, as the Hacker News discussion of the post-mortem (373 points) shows. The market chose Mint, which Intuit bought for $170 million.

History

Wesabe raised about $4.7 million: $700,000 from O’Reilly AlphaTech Ventures in February 2007 and a $4 million round led by Union Square Ventures in June 2007. When it closed, Hedlund said running a service that holds such sensitive data “with shoestring operations and security staff is not acceptable.”

On 1 October 2010 he published “Why Wesabe lost to Mint” on his blog (the original host no longer resolves; it is indexed on Hacker News). As quoted in that thread, his summary was that “Mint focused on making the user do almost no work at all, by automatically editing and categorizing their data,” while Wesabe spent its usability effort in the wrong place. Mint used Yodlee; Wesabe had looked at Yodlee in 2006 and found “the company was crumbling,” so it built its own aggregation. Readers drew the blunt conclusion that users do not pay for privacy with effort.

For ownpurse

Wesabe is the warning in ownpurse’s own positioning. Owner-held credentials and a principled stance did not save a product that made people do the work. ownpurse’s bet is that agents now do that work: they write the connector, categorize, and walk you through the result, so privacy no longer costs effort. Wesabe also shows what a hosted record means at the end: a month’s notice and a download button. A file the owner already holds needs no rescue.

What to borrow: per-source scripts in a plain layout that anyone can extend; one standard shape for messy bank formats; credentials that stay on the owner’s machine. What to avoid: screen-scraping as the primary path when a bank offers a file or an API, and any design where the owner is the categorization engine.