Leash
Security, visibility, and authorization for AI agents
Authorize and monitor your AI agents with policy enforcement, sandboxed execution, and real-time observability—ensuring they operate safely within your defined boundaries.
Authorize your AI agents
AI agents need guardrails. Leash provides the security infrastructure to authorize, monitor, and audit your agents—enforcing policies and boundaries in real-time.
Sandboxed Execution
Every agent runs in an isolated environment with strict resource limits and network boundaries. Contain the blast radius and prevent unauthorized access to your systems.
MCP Authorization
Built-in support for the Model Context Protocol. Authorize which MCP servers your agents can access, monitor tool usage, and enforce permissions across all integrations.
Policy Enforcement
Define what your AI agents can and cannot do using Cedar, a powerful and expressive policy language. Leash enforces policies in real-time, blocking unauthorized actions with fine-grained access controls.
Dev → Deployment
Start local with rapid iteration, then deploy to production with the same security guarantees. Consistent policy enforcement from development through production environments.
Open source. Built for agents.
Leash is available for free, released under the Apache 2 license. Join us in building the future of AI agent security.