Trial safely
Start with normal task tracking and read-only MCP. No card or agent write is required to evaluate fit.
Open quickstartSecurity
The security model assumes agents are powerful actors that need clear scopes, approvals, quotas, and reviewable evidence.
Security review first decision
Yes: evaluate with read-only agent access first, then enable scoped writes only after admins approve permissions and evidence rules.
Start with normal task tracking and read-only MCP. No card or agent write is required to evaluate fit.
Open quickstartAgents use scoped tokens, quotas, audit logs, and human approvals before sensitive work can run.
Review API controlsSecurity whitepaper, DPA, subprocessors, privacy, support, and status pages are public.
Read whitepaperUse sales-assisted review for SSO, retention, custom limits, procurement, and security questions.
Request reviewProduction account verification is gated on 8 evidence-backed checks: SES identity, DKIM DNS, Cognito sender binding, signup delivery, resend delivery, password-reset delivery, bounce and complaint monitoring, and fallback sender rollback.
Source: docs/engineering/cognito-ses-sender-domain.md
The security whitepaper summarizes architecture, Cognito authentication, tenant isolation, audit logs, encryption, backups, CI gates, AWS deployment, and agent-specific approval controls.
Read whitepaperView subprocessors