> About.WhyKarectl
Built from necessity.
Designed for what comes next.
KARECTL is the result of six years of infrastructure and secure data work at Lancashire Teaching Hospitals, driven by the conviction that organisations handling sensitive data should have open, organisation-controlled, AI-capable platforms.
01 // Origin
Starting from a real organisational need.
LTH set out to give researchers controlled access to sensitive patient data, with the aim of reducing infrastructure and vendor costs compared with proprietary solutions and avoiding long-term vendor lock-in. Rather than adopting a closed product, LTH chose to build on open-source, community-backed infrastructure.
02 // Foundation
Six years of infrastructuring.
Over six years, LTH has pursued what Pipek and Wulf call "infrastructuring": the continuous, intentional development of foundational capabilities, including data standardisation, information governance automation, analytical toolchains and changes to operational practice. This foundational work is what makes AI and agentic capabilities possible at scale.
03 // Open foundation
Built on K8TRE.
KARECTL is built on K8TRE, the open-source Trusted Research Environment we co-develop.
04 // Beyond research
One architecture for operational and research AI.
As demand grew, it became clear that the same governed infrastructure could serve live services as well as research. Separate platforms for each would duplicate cost and effort and split governance. KARECTL runs operational AI and research analytics on one architecture, each in its own governed environment, with AI models and agents held to the same controls as people. That is what we mean by AI-conformant, and it is why the platform is working towards CNCF Kubernetes AI Conformance.
05 // Today
A platform for both people and agents.
KARECTL is in production at LTH, on-premises and on Microsoft Azure, supporting operational AI applications and research. The work continues: security threats evolve, governance must adapt to increasingly autonomous systems, and new skills in platform engineering, AI engineering and security operations must be developed. Our aim is a secure foundation for increasingly sophisticated AI workloads in sensitive, regulated settings.