Crowdsourced asteroid discovery (http://www.asterank.com/discover) - only 5 upvotes on HN, but nearly half a million survey images reviewed, with 17,000 potential asteroids marked. Not really working on it anymore.
Free outgoing SMS API (http://textbelt.com/) - not much interest on HN, but about 3M texts sent over the past few years, almost 1000 stars on Github. Requires an hour or two a month for maintenance, responding to issues, etc.
Call Congress (1-884-USA-0234) - single phone number that dials all your representatives one after another. 8 upvotes on HN, but did very well on Reddit and sent over 300 hours of phone calls to Congress in a few days after the Orlando shooting.
Conspiracy theory generator (http://www.verifiedfacts.org/) - Did well on HN with 181 points. About 1M conspiracies generated. On autopilot but still gets organic traffic for ridiculous queries like "snooki illuminati".
Dream logs (http://keepdream.me) - posted 4 years ago but it's a niche tool. 62 subscribers, 2.5k dreams recorded, on autopilot.
Asterank (http://www.asterank.com/) - I submitted parts of this site to Show HN as I added new features. Sold it to Planetary Resources, the asteroid mining company for a small amount.
Meteor showers visualization (http://www.ianww.com/meteor-showers/) - did well on HN, finalist in some Popular Science viz contest.
Dinosaur Pictures database (http://dinosaurpictures.org/) - a few upvotes on HN, about 8k uniques/mo a year later mostly from SEO. This is one of my favorite projects to spend time on.
The sponsors whose content articles get blown up might get pissed with Google and take their ad spend elsewhere...
See here: https://github.com/typpo/textbelt/blob/master/lib/carriers.j...
It seems like textbelt does not know this information (or at least when you sign up or send a text message). Does it try in sequence to send the text to all carriers one by one until one succeeds? If so, I'd imagine textbelt would get blacklisted pretty quickly.
It would be interesting to see how the mapping is done.