In the early days it was quiet but I liked the auto-generated clients. It's really been picking up in the last few months and the hassle-free billing, documentation is great. Augusto is great, he is still personally answering mails when there are issues.
What could be improved on: Better stats for the admin interface, .NET support (I'm aware this has been on their roadmap but customers have repeatedly asked me for this), more payout options like bank transfer would be awesome too.
* Nevermind. Yes they return JSON, I just read the documentation.
For a purely REST API, this might work, but for something like free-to-find services for which I need a quick-api solution to use, it just does not work. The API documentation editor (seriously, stop calling it documentation and call it specs instead) needs an advanced mode so I can use it better. Another thing I found missing was user help text on the apis themselves.
This might take off to become something like programmableweb, but developers won't rush to it with open-arms unless (a few) major api providers start using this.
In the meanwhile, I documented your API with my account, and I would be happy to transfer it to yours if you like. Here it is: http://www.mashape.com/thefosk/heck-yes-markdown
The documentation it is really a documentation, not all the fields are mandatory by the system, except the endpoint URL, and the parameter's name (we need them to auto-generate the client libraries). All the other fields are optional and provided just for the sake of understandability when a third party needs to consume your API.
I'm Stuart, co-founder of chatterbox.co. We've been using Mashape from the early days when they were back in beta, distributing our Sentiment Analysis API (http://www.mashape.com/chatterbox-co/sentiment-analysis-free). For a startup, getting to market as quickly as possible is great.. what you don't want though is a half baked product. Mashape allowed us to focus on building awesome sentiment analysis, so that we didn't have to bother about billing, rate limiting, etc and all the DB infrastructure that runs behind it. Certainly worth checking out, especially if all your stuff uses JSON anyway (which it should!).
I would have thought that the whole point of a custom client would be that it would expose objects and methods that had represented the request and response data for the API.
If I have to interpret and parse the JSON myself then the client code they provide isn't very useful IMHO.
At some point we are going to move to self-describing APIs, documented at source. Check out the http://restfulobjects.org/ spec for ideas on how this will work.
What's are your views here?
This is a really interesting question for me. I keep putting off a side project that I'd love to do because it suffers this exact problem.
I want to believe there's another way to handle this situation.
They app is really well designed and useful without the users there to the API creator.
Therefore, offering a proxy that has enough benefits to developers to make them switch to it's standard seems like a more realistic approach.
So, I truly hope Mashape takes off.
Saying APIs should be standardized is very naive anyway - APIs often need to offer specific platform-specific features.
Edit: back now