How APVISO Works

The Four-Agent Architecture

Every APVISO scan deploys four collaborating AI agents that work together to test your application:

• Recon Agent — maps the attack surface by enumerating subdomains, open ports, technologies, and entry points. It feeds structured intelligence to the other agents.
• Scanner Agent — probes discovered endpoints for vulnerabilities across every OWASP Top 10 category and beyond. It generates context-aware payloads, tests authentication and authorization logic, and chains weaknesses together.
• Lead Agent — coordinates overall strategy. It reviews recon data, prioritises high-value targets, resolves conflicts between agents, and decides when the assessment is thorough enough to wrap up.
• Reporter Agent — documents every confirmed finding with a description, severity rating, evidence (request/response pairs, screenshots), CWE/CVE mappings, and actionable remediation steps. It also produces the final Markdown and PDF report.

Secure Isolation

Each scan runs in a dedicated, isolated environment. Scans are completely separated from one another — one customer's scan can never interfere with another.

Scans have a maximum runtime of three hours. If a scan is still running when the limit is reached, partial results are preserved and the scan is marked as timed out.

Real-Time Results

Agent activity streams to your dashboard in real-time. You can watch recon discoveries, scanner probes, and confirmed findings appear live. There is no need to wait for the scan to finish — you can begin triaging findings as soon as they arrive.

Model Presets

APVISO offers multiple model presets that control the depth and thoroughness of analysis. Higher presets provide deeper reasoning and more creative attack strategies. Your subscription tier determines which presets are available — higher tiers unlock more powerful AI models for deeper analysis where it matters most.