Cognition Box Labs is an specialized technology holding company dedicated to solving the most complex challenges of the digital age. We don’t just build software; we engineer ecosystems of trust, precision and safety.Here are our current holdings across Online Safety, AI and Infrastructure:
Amana Pro is a UK-based platform designed for safe family browsing, featuring AI-powered tools. Amana Pro offers tools to enhance online safety, including desktop computer protection, safe browsing capabilities, and screen time limits. Amana Pro is designed to be easy use, low cost, reliable (leveraging our in-house verification tools), and offers comprehensive online safety features, suitable for many families and communities.
DeepBuild is a UK-based technology firm specializing in the automated creation of formally-verified, safe agentic systems. By fusing a proprietary model-checking algorithm with an intelligent LLM workflow, DeepBuild makes it possible to build and verify LLM based agentic systems. DeepBuild is designed to be more intelligent, faster and more secure than current open-source platforms like OpenClaw (which do not have formal verification and fast simulation baked in), offering guaranteed safety architectures for agentic systems that are suitable for high-stakes industries and complex digital environments.
Petracode is a UK-based technology firm specializing in formal software verification and high-reliability systems engineering. Rather than traditional development, Petracode provides a specialized language and toolset called Petra, which allows businesses to mathematically prove the correctness of their software to eliminate logic errors and security vulnerabilities. By bridging the gap between complex formal methods and mainstream languages like C++ and Java, Petracode enables startups and enterprises in high-stakes sectors—such as fintech, healthcare, cybersecurity and AI—to build mission-critical digital solutions with a foundation of absolute mathematical certainty and high-performance reliability.
Agentic Trader is a proprietary algorithmic trading system which connects to Saxo Bank for FX and Polymarket prediction markets. Rather than relying on black-box heuristics, Agentic Trader utilizes a proprietary framework of deterministic machine learning and verified agentic systems to navigate the complexities of Foreign Exchange and Prediction Markets. By integrating mathematically rigorous models with adaptive execution logic, the platform enables institutional-grade participants to deploy strategies that are not only high-performance but also computationally verifiable. By bridging the gap between predictive sentiment and automated execution, Agentic Trader empowers traders in high-stakes environments to capitalize on market inefficiencies with a foundation of algorithmic precision and systemic reliability.