Accountability for AI Agents
Identity, policy, pre-action authorization, and proof for agentic work. APort checks actions before tools run and records signed decisions for audit.
Agent Passports
Issue Open Agent Passport records for agent identity, ownership, capabilities, limits, regions, and status.
- Self-serve passport issuance
- Hosted or local passport modes
- Framework-tailored presets
Audit Trail
Tamper-evident Ed25519 signed decision records that support audit evidence and incident review workflows.
- Signed decision records
- Multi-format export
- Searchable policy reasons
Pre-Action Authorization
Evaluate shell, file, MCP, web, repository, finance, messaging, and export actions before tools run.
- 21 open policy packs
- Data export controls
- Sensitive command and file checks
GitHub Repository Guard
Use GitHub OIDC and repository-scoped OAP passports to verify pull requests and protected pushes before merge.
- No broad APort API key in GitHub secrets
- Protected path findings
- Signed merge decisions
Runtime Guardrails
One-command guardrails for Claude Code, Cursor, OpenClaw, LangChain, CrewAI, MCP, and custom agent workflows.
- Claude Code and Cursor hooks
- OpenClaw plugin setup
- MCP and framework integrations
Multi-Region Options
Hosted regional routing for passport, policy, and audit data across supported deployment regions.
- US, EU, CA regions
- Tenant-aware routing
- Enterprise residency options
Enterprise Rollout
Hosted and local enforcement modes, org controls, delegated setup keys, and IT rollout scripts.
- Hosted and local modes
- Device deployment scripts
- Dedicated support by agreement
Built Around Verifiable Evidence
Ready to Get Started?
Start with a passport, wire one guardrail, and turn agent actions into signed decisions your team can review.