From Dropbox to Google Docs, cloud applications are increasingly becoming a part of business users’ everyday productivity toolbox. Hojoki – taken from the name of an old Japanese book about the way of life and the flow of events – aims to aggregate those cloud apps into a Facebook-like activity feed, underpinned by semantic technology.

“We really think the real pain point up to now is that work and communication about work are separated,” says Martin Böhringer, co-founder and CEO. “We integrate the productivity apps that people already are using so that they show up in the stream” where people are communicating.

Currently the service supports integrations – either in public or private beta mode – for Google Calendar, Google Docs, Twitter, Dropbox, Evernote, Delicious, Github, CloudApp, Zendesk, and a few more, with roadmap plans to add Google Reader, Google Contacts, Lighthouse and Mendeley. The goal is to deliver about three new integrations each month. Also, says Böhringer, there’s consideration being given to using the Atom Activity Streams RDF mapping Vocabulary Specification standard so that users can write their own integrations for social networking activity streams.

With Hojoki, users can view their collected cloud app activity streams for types (such as codes, bookmarks, conversations or files and folders) or persons, and sort accordingly; and build groups for different projects to which they can automatically share activities, create updates, and get feedback and notifications from other collaborators.  Its broad integration plans to encompass multiple cloud tools, each with its own different concepts and data types, “is why we chose a semantic data structure,” says Böhringer. The heart of the system is a generic activity store based on RDF. “We have the task to integrate lots of different kinds of services but we need some kind of structure.”

The activity stores in Hojoki, a J2EE application, include all activities of a group that flow in from cloud apps. Activity stores form a distributed network consolidated by the API layer, the company says, and the stores are based on RDF to allow complex search and analysis as well as seamless integration with existing semantic data. “Each project, Böhringer explains, is its own database, its own RDF store. “So, if you are involved in five projects you have access to five data stores. Those data stores can be everywhere but the special trick in Hojoki is that the UI combines all of them. It’s basically a federated system on the back end that’s combined in the front end, where the stream is generated by JavaScript.”

Think of semantic technology as enabling the system to classify all tasks that flowing in as such and providing the same user interface into them. “If you look on Hojoki for tasks you see tasks from all your systems,” he says.  Also enabled by the semantic technology is the ability to define very rich rules of what kind of information can be connected to a project, he says. The company expects in the future to leverage its semantic heritage to provide even richer categorization, such as helping users seamlessly discover which other documents are related to the one they’ve just viewed, who are its collaborators, and so on.

The company recently raised $620,000 with investor Kizoo. Right now its services are all free, though premium accounts are planned for the future.