Haven OnDemand, which is hosted on Microsoft’s Azure platform, provides developers with APIs and services for building data-rich applications. The roughly 60 APIs currently included in the service include features like sentiment analysis, speech recognition services, trend analysis and text classification services. Haven also includes a prediction and recommendation API that developers can train for their specific use cases. The service also includes an image recognition feature, but it’s still rather rudimentary for now and only recognizes a set of corporate logos in images. The company tells me it plans to invest more in this area going forward, though.
As Colin Mahony, the company’s senior VP and general manager for HPE Big Data, told me, the company is targeting both enterprise developers and startups with this service. HPE wants to give them a flexible platform that can grow with them as their needs scale up.
“Developers are good at what they do, but they are not statisticians and mathematicians,” he told me. “They want us to take a lot of the statistics, heuristics, and machine learning, and build it into things that do advanced text analytics, or basic format conversion, facial detection, etc.”
Now that it is commercially available, Haven OnDemand will offer a free tier that limits the number of API calls and storage developers can use. Paid plans start at $10 per month. For the first three months, HPE is also offering an additional 50,000 API units (which roughly equate to API calls) for all paying users.
The company says 12,750 developers registered for the service since 2014 and they currently generate “millions of API calls per week.” This has allowed HPE to fine tune its service to the point where it is now confident that it can offer SLA agreements to enterprises.
To show off the power of the platform, HPE also offers enterprises a search platform (Haven Search OnDemand) that uses Haven’s APIs.