Card.io was started by a couple of former AdMob engineers and product managers to solve what they perceived to be a major problem in so-called "m-commerce," accepting plastic payments. There are a ton of companies competing in the mobile payments segment but there's nobody right now doing what Card.io is doing.
Card.io uses the smartphone camera to enable consumers to get credit card information into the phone efficiently without keying in 16 digits. Their SDK is intended for developers and publishers who want to sell things via their mobile sites. Card.io doesn't compete with Google Wallet or Square; it's not merchant-facing or consumer-facing. It solves the problem of abandonment when consumers confront the proposition of entering credit card numbers for the first time into mobile websites.
One of the major reasons why Amazon is "killing it" in mobile commerce is because the company already has your credit card on file so the friction involved in completing the transaction on a smartphone is eliminated. They also have a trusted brand and have put huge effort into optimizing for mobile.
I asked CEO Mike Mettler about potential security concerns regarding taking and transmitting pictures of credit cards via smartphones. He said they don't store credit card information and are very specific and careful about security in their messaging. He said that they anticipated this problem but so far haven't seen the concern materialize.
In my view this is a potential "game changer" for e-tailers that don't or can't have per-existing direct relationships with consumers (stored credit cards). I would expect any retailer who sells goods online to integrate this capability into their mobile sites. I would also expect that Card.io will be an acquisition target in 2012.
The company currently charges a per-scan fee to developers but will probably be moving to a percentage-based transaction model. Below is a link to a video that demonstrates the process.