Lab
The lab is where our products take shape before release. We build AI tools across endpoint security, investment guidance, and automated software testing. Same team, same standard. Some are approaching early access.
Private beta
Continker AV
Endpoint threat detection built on a custom transformer architecture. Designed to reason about code, not just match against known signatures.
Continker AV disassembles ELF binaries into a crafted intermediate representation that exposes program structure. A purpose-built transformer relates and weighs across tokens. It reads how a binary is organized, reasons about intent, and detects novel malware that has never been catalogued. A second tier fuses detection with a reasoning model for deeper analysis and human-readable insight.
It runs wherever Linux runs. IoT devices, routers, single-board computers, servers, containers, workstations. Light enough for constrained edge hardware, thorough enough for production infrastructure. Windows support is in development.
European-hosted, privacy-first. No bulk telemetry. Built for enterprise security teams, managed service providers, and anyone running Linux at scale.
Coming soon
Belero
AI investment assistant for self-directed investors. Systematic analysis delivered through conversation.
Belero classifies the macroeconomic cycle using GDP trends, inflation, unemployment, monetary policy, and yield curve data. It scores sectors on four components: cycle fit, relative performance, valuation gap, and trend alignment. Individual stocks pass through a multi-step screening pipeline covering technicals, financials, news impact, and benchmark comparison. A scheduled scanner runs the full analysis daily and delivers a morning briefing.
The interface is conversational. Ask about a stock, request a sector overview, set a price watch. Belero responds with structured analysis in natural language through configurable personas that give each conversation its own character and tone. Your funds stay where they are. Belero analyzes and advises, you invest on the platform of your choice.
Built on public market data and central bank APIs.
In development
Agentic QA
AI-driven usability testing that simulates how real people experience your product. Scenario-based, persona-driven, repeatable at scale.
The agent interprets your interface visually and follows scenarios shaped by persona characteristics. An elderly person on a tablet navigating a map widget. A first-time visitor searching for a support page. A technically limited user completing a checkout form. When the persona gets stuck, the agent adapts. It tries alternative approaches, different navigation paths, workarounds. It behaves the way the instructed persona would behave.
Each scenario runs repeatedly with nuanced variations in persona and instruction. Failures, retries, and alternative solutions are counted across runs. A second AI layer analyzes the full set of results and produces a structured assessment: what worked, what failed, and what to improve.
Interested in early access or a partnership? Get in touch →