Automated AI scanning across five agent-readability dimensions. Score tracking, instant alerts — so you catch problems before your customers do.
Agent Analytics is free to use during beta · Paid tiers open July 2026
Add any URL. COVEN begins scanning immediately.
AI analyses five agent-readability dimensions on your chosen schedule.
Score drops trigger email or webhook within minutes.
What's included
Structure, Citability, Discoverability, Agent Interface, and Transactability — all scored on every scan. Agent Correlation reported as a research signal.
No jargon. When something drops, we tell you exactly what changed and why it matters.
Track how your site trends over time. See the impact of every change you make.
Score changes POST to your endpoint with full diff data. Build your own response logic.
Add COVEN as an MCP tool. AI agents can register and query monitored sites natively.
Unreachable sites trigger an immediate alert. Know before your customers do.
Add competitor URLs alongside your own sites. See how your scores compare side by side — no alerts, just intelligence.
For Developers & Agents
COVEN Monitoring is API-first. Register sites, receive webhooks, and integrate with any AI agent or automation pipeline.
Every scan returns machine-readable scores, dimension breakdowns, and issue lists.
Full spec at /openapi.json — compatible with any HTTP client or agent framework.
Beta
Agent Analytics is free to use right now during our beta. Paid subscription plans — higher limits, team access, and priority support — open after beta in July 2026. Leave your email and we will let you know.
Ready to start now? Open Agent Analytics — free during beta, no card required. Or try the free one-off scan.
FAQ
The same five scored agent-readability dimensions as our citability scanner: Structure (25%), Citability (25%), Discoverability (15%), Agent Interface (20%), and Transactability (15%). Every scan scores each dimension out of 100. Agent Correlation is reported alongside as a research signal at 0% weight.
Within minutes of a scan completing. Any score drop of more than 10 points will trigger an immediate alert. Monitoring opens after our beta period.
Yes — fully API-accessible and MCP-compatible. Agents can register sites, query scan history, and receive webhook alerts. See the developer docs for the full reference.
Unreachable sites trigger a site_down alert immediately, separate from the scheduled scan cycle.
Run a free agent-readiness scan while monitoring is in development.
Run a free scan →