Roadmap

What AgentGate does today, what is in beta and what is planned. Register interest in a planned capability.

• Browser check: An invisible check that the page really ran in a browser. People see a prompt only when something looks off.
• Server verification: Redeem each token once on your backend with POST /v1/siteverify.
• Gateway: Check requests before they reach your app, with nginx auth_request or middleware.
• Rules: WAF-style ordered rules over labels, IP sets, ASN, headers and rates, plus managed rule groups.
• Threat intelligence: ASN and country, cloud ranges, Tor exits, verified crawlers and decaying IP reputation.
• Signed agents: Verify Web Bot Auth signatures and give each agent its own lane and rate limits.
• Agent policy: Per-agent actions and quotas, 402 pricing signals and a handoff to a person.
• Dashboard: Every decision with its labels, matched rule, scores and model version.
• Customer console: Add sites, copy keys and snippets, check the install, then switch from monitor to enforce.
• Usage metering and billing (planned): A usage ledger, plans and invoices for self-service sign-up.
• Teams and roles (planned): Invite teammates to a console account and choose who can change what.
• Structured rule editing (planned): Edit routes, actions and rules in forms, with JSON kept as the advanced view.
• Challenge funnel analytics (planned): Follow checks from requested to solved, verified and allowed.
• Pay-per-crawl settlement (planned): Settle what agents owe; today AgentGate only signals prices.
• Multi-region hosted service (planned): Run the hosted service in more than one region, with shared state.
• Real-user model evaluation (planned): Measure detectors on consented real traffic, including assistive technology, before they enforce.