For about two years, I’ve maintained a Ruby library called EasyTranslate for working with the Google Translate API. Google Translate has been (like other Google APIs) free to use withing a certain range.

Recently, they’ve deprecated (read: completely turned off) access to v1, due to substantial economic burden caused by extensive abuse . That’s really a shame. The larger shame is that there is no free usage tier for v2 of the API. Even to santiy check my API client (that I don’t even use anymore), I need to pay.

This reminds me of an article I read yesterday on HackerNews (and linked by Kenny Katzgrau) about how one developer is fed up with freemium as a model due to lack of conversion from free to paying customers. I don’t really agree with that article for a few reasons:

  • The timeframe of his freemium strategy was too short to build the devoted user base that freemium depends upon
  • The product he was working with is (however well done it may be) not something that users will build loyalty to, but rather a seasonal site of a sort that most people would reasonably expect to find for free
  • If you expect a high number of people to pay, just charge for the service and accept the lower virality along with making the bar of entry higher. I'd argue if he hadn't had a freemium model in place at all, his overally hit count and paying customer count would have been even lower.

But the article does frame a discussion for deeper problems with freemium APIs like Google s. Automated abuse can quickly ruin a great service, but I have to think there must be a middle ground between non-keyed API access (as before), and no free-usage tier. What do you think?

NOTE: Last night I largely reorganized and re-spec ed easy_translate to work solely off of the examples in the API docs. Its released now as v0.3.0.