# AgentRail

> AgentRail is the independent runtime control layer for enterprise AI agent actions. It inspects every AI action before execution and lets teams allow, require human approval, or block — across HubSpot, GitHub, Okta, Salesforce, Slack, Jira and internal workflows. AgentRail is EU AI Act ready (Article 12 logging + Article 14 human oversight).

AgentRail is built for security, compliance and platform teams who need to govern autonomous AI agents in production. The platform spans four runtime capabilities:

- **Discovery** — map every agent and every action across your systems
- **Risk scoring** — deterministic per-action risk scoring
- **Policy enforcement** — declarative policy as code, allow / approve / block in real time
- **Immutable audit** — per-action evidence ready for the EU AI Act

Key facts:
- Vendor-, model- and framework-agnostic (works with LangChain, LangGraph, CrewAI, custom agents)
- Native connectors: HubSpot, GitHub, Okta, Salesforce, Slack, Jira and more
- EU AI Act: Article 12 (Logging) + Article 14 (Human oversight) covered out of the box
- Founded: 2025
- Website: https://agent-rail.dev

## Core pages

- [Home](https://agent-rail.dev/): Product overview and runtime governance pitch
- [Product](https://agent-rail.dev/product): Discovery, risk scoring, policy, audit trail
- [Railgun](https://agent-rail.dev/railgun): Real-time action gating engine
- [Runtimes](https://agent-rail.dev/runtimes): Supported agent runtimes and frameworks
- [Developers](https://agent-rail.dev/developers): SDKs, policy as code, integration guides
- [Security](https://agent-rail.dev/security): Security posture, SOC 2, ISO 27001
- [EU AI Act compliance](https://agent-rail.dev/ai-act-compliance): Article 12 + Article 14 mapping
- [What is AI agent governance?](https://agent-rail.dev/ai-agent-governance): Pillar guide and FAQ
- [About AgentRail](https://agent-rail.dev/about): Company, mission, team

## Comparisons

- [AgentRail vs Lasso Security](https://agent-rail.dev/vs/lasso-security)
- [AgentRail vs CalypsoAI](https://agent-rail.dev/vs/calypsoai)
- [AgentRail vs Lakera](https://agent-rail.dev/vs/lakera)

## Blog

- [Why runtime governance matters](https://agent-rail.dev/blog/why-runtime-governance-matters)
- [Discovery before governance](https://agent-rail.dev/blog/discovery-before-governance)
- [Risk scoring AI agents](https://agent-rail.dev/blog/risk-scoring-ai-agents)
- [Policy as code for AI governance](https://agent-rail.dev/blog/policy-as-code-ai-governance)
- [Human-in-the-loop AI agents](https://agent-rail.dev/blog/human-in-the-loop-ai-agents)
- [EU AI Act: what it really changes for enterprises](https://agent-rail.dev/blog/ai-act-what-it-really-changes-for-enterprises)

## Contact

- LinkedIn: https://www.linkedin.com/company/agent-rail/
- X / Twitter: https://x.com/AgentRail_dev
- Discord: https://discord.gg/YFkdzCFU