UNISI Roadmap 

UNISI is an extremely easy-to-maintain system, and generating modules, screens, and interfaces to third-party services by LLMs is possible in any time. The system designer will design the entire system from scratch, demonstrating interfaces and conducting a clarifying dialogue with the user, such as yes, no, what do you choose from this, relying on the capabilities of UNISI infrastructure and generating binding code and descriptions. As a result, you'll get a ready-to-use system or a set of interconnected systems.

The UNISI GUI enables drag-and-drop functionality to transfer data from one system to another for further processing or connecting. The sequence of such transfers and execution commands can be saved in Units and utilized in any UNISI system. ​

We eagerly await the release of Mojo 1.0, which is expected to replace Python in the future for high-load, computationally complex, and most AI tasks. The interfaces will be the same, but all code will be rewritten, taking into account the capabilities and advantages of Mojo. Mojo and Python will be equally supported in UNISI-Mojo.

UNIDB will replace existing kuzudb for database management. Existing databases are not optimized for working with large, distributed networks of objects, require substantial overhead, and operate significantly slower than an optimized solution designed for the UNISI data model. UNIDB will increase performance by orders of magnitude, enable data fragmentation, and facilitate data exchange without complexities and headache.

The new GUI subsystem will allow customization of styles, sizes, and fonts of interface elements and will increasingly utilize animations for transitions. The primary feature is the user space seamlessly interfacing with multiple UNISI systems concurrently, as though interacting with a single unified system.