Who we are
Safetrek.ai emerged from the operational experience of BA Clearance GmbH, a company specialized in explosive ordnance disposal in challenging environments and headquartered in Tautenhain, Thuringia. From hands-on field operations arose the need to assess dangerous situations faster, more accurately, and in a more data-driven way.
Our team consists of experienced former international special forces personnel and ordnance disposal experts. We are dedicated to supporting complex operations and high-risk environments where conventional ordnance disposal capabilities are limited. Backed by a reliable network and many years of operational and training experience, we are able to identify, classify, assess, and dispose of ordnance both on land and in maritime environments.
Having not only deep expertise in this field, but also years of experience in additional sectors such as production and corporate environments, our portfolio has continuously expanded, and with it, our scope and presence.
What we do
We started by solving one of the most complex detection problems in the world: identifying explosive remnants of war from the air, from hand-held devices, and directly from the ground. Every line of code, every model, and every workflow was built and hardened under real operational constraints, where reliability is not optional but essential.
This foundation shaped our approach to AI from the beginning. What was developed for high-risk environments has become a broader capability for perception, decision support, and operational awareness. Today, we apply the same rigor to industrial quality, process optimization, and safety-critical environments. Our platform delivers real-time AI perception wherever operations take place - on smartphones, autonomous systems, industrial robots, humanoid co-workers, or fixed edge infrastructure. It combines fast, on-device detection with deeper contextual scene understanding when complexity demands it.
Because the engineering bar in defense is the right bar. Everywhere.
Domains We Operate In
Today, safetrek.ai operates at the intersection of defense and industrial autonomy. We bring defense-grade rigor into environments where autonomy is rapidly scaling: factories, logistics networks, mobility systems, and critical infrastructure. In these settings, complexity is increasing faster than traditional safety frameworks can adapt, creating a clear gap between system capability and system assurance.
Our approach is to close that gap. We translate the principles proven in defense, continuous situational awareness, structured decision-making, and resilient system design, into a unified safety and intelligence layer for modern autonomous operations. This ensures that systems are not only high-performing, but also observable, explainable, and controllable under real world conditions. Rather than treating safety as an external constraint, we integrate it as a core system capability. This allows organizations to deploy robotics and AI with confidence, knowing that behavior can be monitored, risks can be anticipated, and responses can be executed in real time.
Safetrek.ai exists at the intersection of autonomy and accountability. Where systems are no longer just built to operate, but to operate safely, reliably, and transparently at scale.
Why safetrek.ai
One perception engine, built to a defense-grade standard – deployed wherever the mission lives.
One platform, every deployment surface
The same perception engine runs on smartphones, UGVs, UAVs, AMRs, KUKA cells, humanoids and fixed edge nodes - delivered as a containerised service, as a native application, or embedded directly via SDK.
Strict data isolation
Customer data stays in the customer's pipeline.
Training data, inference logs and IP are separated by architecture, not by policy.
Edge-first, cloud-optional
Inference happens on the device, with optional sync to the backend. Works under denied or restricted connectivity.
Forensically auditable
Every safety-critical event is timestamped, signed and traceable.
Accident reconstruction, EOD after-action and QM root-cause analysis all use the same evidence chain.
Ultimate speed of adoption
New defect classes, new ordnance types, new hazards - integrated within days, not months.
© 2025 safetrek.ai GmbH - All rights reserved