Teleport makes Beams agent runtime generally available

Since November 9, 2026, Teleport Beams has been generally available as a trusted runtime for securing AI agents in production infrastructure, following a public beta that started in March 2026. The product addresses a barrier that has kept most enterprises from linking agents to live systems.

The company says a mistake or hijacked agent has a small, visible blast radius, which it presents as what allows security teams to approve agents and move them from prototype to production. Traditional sandboxes, which focus on virtual machine isolation and credential management, are described as insufficient when an error is costly. Beams is built on the Teleport Infrastructure Identity Platform, which operates in production at more than 700 enterprises.

Beams applies four layers of control that govern both individual and swarm behavior. The identity layer gives each runtime a delegated identity from a user or another agent, with agents starting at zero standing privileges. The access layer connects agents to cloud accounts, Kubernetes clusters, PostgreSQL databases, and MCP servers without secrets or passwords in the agent environment, authenticating through IAM, Kerberos, X.509, and other identity-based methods.

Agents submit execution plans for review by a human or an LLM judge, and a separate runtime carries out approved plans. The swarm layer flags goal hijacking and emergent coordination between agents, scoring each session against the MITRE ATT&CK framework. Agents run in ephemeral micro-VMs with virtualized networking that controls egress, ingress, and DNS, and base images are rebuilt regularly to reduce CVEs.

“Most enterprises have not connected agents to their production infrastructure. They are blocked by a lack of delegated identity, strong access control without secrets and with full audit and traceability of intent to action. When the cost of an error is high, traditional sandboxes don’t cut it. Beams brings these missing capabilities to the market,” said Ev Kontsevoy, CEO and co-founder of Teleport.

“Platforms that build these into the runtime are helping move agentic AI from prototype to production,” said Grace Trinidad, Research Director for AI Security & Trust at IDC.