Concepts & Glossary
The shared vocabulary behind Noxtica. Every term the product and the homepage use resolves to a real definition here — what it means, how Noxtica does it today, and where to read the deeper technical doc.
Noxtica is an Agentic Security & Intelligence Platform. Two halves: an intelligence layer that reads who and what is on your site — device, agent, browser signals — and a security layer that acts on what it reads, in the browser, under policy you control. Running through both is an agentic spine: an operator-facing AI assistant that drives the console, a read-only integration surface so your own agents can read Noxtica, and a defensive registry that governs which AI agents you trust. None of it is magic. The reads are calibrated and explainable; the actions are policies you write.
The Intelligence pillar
Reading who and what is on the page.
- Know Your Device (KYD) — Sealed Runtime device intelligence: cross-browser signals condensed into a stable, tamper-resistant fingerprint, a persistent device ID, and a risk score.
- Know Your Agent (KYA) — a defensive allow/deny registry that governs which AI agents and bots you trust, integrated with Web Bot Auth verification.
- Browser Intelligence — the signal-collection layer: the categories of evidence the Sealed Runtime gathers before any decision is made.
The Security pillar
Acting on what was read.
- Browser Security — client-side enforcement: challenge and block directives, per-domain policy, shadow mode, and the operator-configurable risk-action engine.
The Agentic spine
Three honest meanings of “agentic”, all shipped.
- Agentic Security — the umbrella concept: we operate agentically, you integrate agentically, and we police the agentic web. How KYA, MCP, the assistant, and the risk-action engine fit together.
- MCP Integration — a read-only Model Context Protocol server so your external AI agents can read Noxtica over JSON-RPC, with scoped, rate-limited, audited tokens. Opt-in per tenant.
- AI Assistant — the built-in, server-side assistant that helps operators run the console, with per-tenant budget caps, audit logging, and read-only tools in Phase 1.
Calibration
The principle under everything.
- Calibration — why Noxtica returns calibrated risk reads, not verdicts, and where the calibration roadmap is headed.
Related reading
- Engineering principles — the operating constraints behind every surface.
- Why calibration, not verdicts — the philosophy under the whole platform.
- Detection signals — the probe-level detail behind the intelligence layer.
- Integration flow — script tag to risk-scored decision, end to end.