A Swiss hardware maker and an Italian AI specialist just announced a partnership designed to remove friction from deploying artificial intelligence directly on edge devices, without relying on cloud connectivity. Enclustra, a provider of field-programmable gate array (FPGA) and system-on-chip (SoC) modules, has integrated MakarenaLabs' MuseBox orchestration framework, rebranded as Lira when running on Enclustra hardware, to give developers a tested path from silicon to working AI pipelines.

What Problem Does This Partnership Actually Solve?

The core challenge in edge AI deployment is fragmentation. Engineers building real-time, on-device intelligence systems have historically faced a patchwork of hardware options, software frameworks, and integration hurdles. They must connect live data sources, accelerate inference using specialized hardware, and deploy results to dashboards or physical systems, all while keeping everything running locally without mandatory cloud dependency. This is especially critical in industries where connectivity cannot be guaranteed or where latency matters more than cloud processing power.

Enclustra brings two decades of experience building FPGA and SoC modules trusted in industrial, aerospace, and research environments. MakarenaLabs contributes expertise in edge AI orchestration and hardware acceleration. Together, they're positioning Lira as a production-ready platform that reduces uncertainty and gives developers a clear deployment path.

"At Enclustra, we've spent two decades building FPGA and SoC modules that engineering teams trust to work reliably in the field. Partnering with MakarenaLabs extends that reliability to the AI layer, giving our customers a tested path from silicon to a working edge AI pipeline, on hardware they already trust," said Philipp Baechtold, CEO of Enclustra.

Philipp Baechtold, CEO of Enclustra

What Capabilities Does Lira Bring Out of the Box?

Rather than requiring developers to build AI pipelines from scratch, Lira arrives with pre-built, real-time capabilities including face detection and recognition, object detection and depth estimation, and hand and face landmarking. The platform supports real-time video, image, and audio processing, WebSocket integration for external system communication, and an optional no-code graphical user interface (GUI) that lets non-specialists build and orchestrate complex AI applications.

Because Lira runs across Enclustra's SoC, MPSoC (multiprocessor system-on-chip), and MLSoC (machine learning system-on-chip) modules, teams can scale from compact standalone deployments to high-performance, multi-accelerator pipelines using the same framework. This consistency matters in production environments where retraining teams on new tools introduces risk and delays.

How to Deploy Edge AI Using This New Platform

• Select Your Hardware Module: Choose from Enclustra's SoC, MPSoC, or MLSoC System-on-Modules based on your performance and power requirements, knowing that Lira will run consistently across all options.
• Use Pre-Built AI Models: Leverage Lira's out-of-the-box capabilities for computer vision tasks like face detection, object recognition, and depth estimation rather than training models from scratch.
• Build Your Pipeline with the No-Code GUI: Connect data sources, configure inference acceleration, and route results to dashboards or actuators without writing custom integration code.
• Deploy Entirely at the Edge: Run your complete AI pipeline locally on the hardware module, eliminating cloud dependency and reducing latency for real-time decision-making.

Which Industries Stand to Benefit Most?

The partnership targets five primary sectors where local, real-time decision-making is non-negotiable and cloud connectivity cannot always be guaranteed. Smart city infrastructure, such as traffic management and public safety systems, benefits from instant processing without network delays. Medical and healthcare applications, including diagnostic imaging and patient monitoring, require low-latency inference and often operate in environments with limited connectivity. Industrial inspection systems, used in manufacturing and quality control, need real-time defect detection without sending sensitive data to external servers. Defence applications demand secure, offline-capable AI processing. Robotics systems require instant perception and decision-making to respond safely to dynamic environments.

"We saw this partnership as a combination of Italian tailoring and Swiss watchmaking. Enclustra brings the precision, reliability, and engineering excellence of Swiss hardware; at MakarenaLabs, we've tailored the AI layer around it to fit real market needs," explained Enrico Giordano, CEO and CTO of MakarenaLabs.

Enrico Giordano, CEO and CTO of MakarenaLabs

When Will Lira Be Available?

Lira, powered by MakarenaLabs' MuseBox technology, will be available soon across Enclustra's SoC, MPSoC, and MLSoC System-on-Modules. The partnership represents a shift in how edge AI infrastructure is being packaged, moving away from point solutions toward integrated platforms that combine trusted hardware with proven software orchestration.

The announcement underscores a broader industry trend: as on-device AI becomes more critical to real-world applications, the companies winning market share are those that simplify the entire stack, from silicon selection through deployment and operation. By combining Enclustra's hardware reliability with MakarenaLabs' AI orchestration expertise, the partnership removes a significant barrier to adoption for engineering teams building the next generation of intelligent edge systems.