Teleport gives engineers, machines, workloads and AI agents cryptographic identities and uses them to control access to infrastructure: servers, Kubernetes clusters, databases, internal applications, cloud consoles and MCP servers. It is built for organisations that want to stop relying on passwords, shared SSH keys, VPNs and long-lived API tokens.
Instead of static secrets, Teleport issues short-lived certificates at the moment of access, backed by a hardware root of trust, and records what each identity does.
At A Glance
| Vendor | Teleport (Gravitational Inc.) |
|---|---|
| Category | Infrastructure identity, zero trust access, machine and workload identity |
| Identities covered | Human users, machines, workloads and AI agents |
| Deployment | Teleport Enterprise Cloud, self-hosted Teleport Enterprise, or Community Edition |
| Agentic AI | Secure MCP access, Beams trusted runtime for agents |
| Certifications | SOC 2 Type II, ISO 27001, HIPAA Security Rule compliance |
| Customers | Nasdaq, DoorDash, Accenture, GitLab, Discord |
What Teleport Does
Teleport targets identity fragmentation in infrastructure. Most teams run one access system for SSH, another for Kubernetes, others for databases and cloud consoles, and lean on shared secrets and standing privileges to glue them together. AI agents add a new class of actor to that mix.
Teleport replaces those separate systems with one identity layer. Every person, bot, service and agent gets a cryptographic identity, access is granted for as long as it is needed and then expires, and every session lands in one audit trail. The products below all sit on that layer, so the same roles, approvals and logs apply whichever kind of identity is connecting and whatever it is connecting to.
Products
Zero Trust Access
Zero Trust Access lets platform and security teams retire shared SSH keys, VPNs and standing passwords, giving engineers and automation passwordless access to infrastructure through a single entry point. Each connection uses a short-lived certificate tied to the user’s single sign-on and MFA, and role-based access control decides what they can reach. Sessions are recorded and actions written to a structured audit log. Device trust can limit access to registered company devices, and hardware security keys give phishing-resistant login.
- Linux, OpenSSH and Windows servers over SSH and RDP, with or without Active Directory
- Kubernetes clusters and the applications running in them
- Databases
- Internal web and TCP apps, cloud APIs, cloud consoles and Git repositories
Identity Governance
Temporary access has a habit of becoming permanent, and audits then need a record of who held which role and why. Identity Governance addresses both. Access Requests let users ask for an elevated role or a specific resource just in time, approvals can be automated with RBAC, ABAC or contextual rules, and the access expires on its own. Access Lists group users and grant roles with periodic reviews, Access Monitoring points out excessive or unused privilege, and a compromised identity can be locked straight away.
Machine & Workload Identity
Pipelines, scripts and services often run on long-lived API keys that leak or never get rotated. Machine & Workload Identity gives these non-human actors, and AI agents, short-lived identities in place of stored keys, in two modes.
- Machine ID issues Teleport credentials so bots can reach Teleport-protected resources, for example running Ansible over SSH or deploying to Kubernetes from a pipeline
- Workload Identity issues SPIFFE-compatible X.509 and JWT SVIDs for service-to-service mutual TLS and for authenticating to third-party APIs and clouds, including Google Cloud Workload Identity Federation
Identity Security
Identity Security gives security teams a view of every access path into their infrastructure. Access Graph maps the paths humans, machines and agents can take to critical resources, including access granted outside Teleport through tools such as Okta, AWS and GitHub, and highlights standing privilege and unmanaged SSH keys. Crown Jewels marks the most sensitive assets for priority alerting, AI summaries describe what happened in each session, and Risk Scoring ranks SSH, Kubernetes and database sessions and maps actions to MITRE ATT&CK.
Beams
Beams is a trusted runtime for teams putting AI agents into production and needing to contain what each agent can reach and do. It runs agentic workloads in ephemeral micro-VMs, gives each agent an attestable identity and enforces its operational, execution and communication boundaries. Agents reach cloud infrastructure without secrets or passwords, and users can delegate a subset of their own access to an agent.
How It Works
A Teleport deployment is a cluster. Infrastructure is enrolled into it, and trusted clusters can federate access across several clusters.
tsh client
| Component | What it does |
|---|---|
| Auth Service | The cluster’s certificate authority. It verifies identities, signs short-lived host and client certificates, enforces RBAC policy and stores audit events and session recordings. |
| Proxy Service | The single public entry point for people and machines. It serves the web UI, terminates connections and routes traffic to the right resource over TLS. |
| Teleport agents | Run beside protected resources and handle SSH, Kubernetes, database, application and desktop traffic. They dial out to the Proxy over reverse tunnels, so resources never accept inbound connections from the internet. A Discovery Service finds new cloud resources such as EC2 instances, EKS clusters and RDS databases and enrols them automatically. |
| tbot | A lightweight agent that sits beside each workload, proves its identity to the cluster and keeps its certificates renewed. Credentials are written to disk or Kubernetes secrets, or served over a local SPIFFE Workload API socket, and on Kubernetes it can run as a daemon set for every pod. |
| Join methods | Workloads prove who they are using the platform they already run on rather than a shared secret: AWS IAM, Azure, Google Cloud, Kubernetes service accounts, GitHub Actions, GitLab CI, CircleCI, Terraform Cloud, Spacelift, TPM for on-premises hardware, or a bound keypair. Workload attestation lets the SPIFFE ID reflect the calling process, such as a namespace and service account, a Unix user or a systemd unit. |
Teleport Enterprise Cloud runs the Auth and Proxy services for the customer; self-hosted customers run the whole cluster. People connect through the web UI, the Teleport Connect desktop app or the tsh command-line client.
Identities And Systems Covered
| Human users | Engineers and administrators, synced from the organisation’s identity provider |
|---|---|
| Machines and workloads | CI/CD jobs, services and bots, governed as Bots under the same roles as people |
| AI agents | Treated as distinct identities with short-lived credentials, alongside LLM tools, MCP tools and digital twins |
| Servers and desktops | Linux, OpenSSH and Windows servers, Windows desktops, and Linux developer desktops (added August 2026) |
| Cloud | AWS, Microsoft Azure and Google Cloud, including EKS, AKS and GKE |
| Databases | 26 types, including PostgreSQL, MySQL, MongoDB, Redis, Microsoft SQL Server, Snowflake, Oracle, Amazon Aurora, DynamoDB, DocumentDB, Cassandra, CockroachDB, MariaDB and ScyllaDB |
| Code and AI resources | Git repositories, model weights and model registries, and MCP servers |
Agentic AI Capabilities
Teleport handles AI agents as a fourth identity type next to humans, machines and workloads, with the same cryptographic identity, least privilege and audit.
Secure MCP Access
Since Teleport 18, the platform can sit between MCP clients such as Claude Desktop or Cursor and MCP servers. Users and agents connect with their own identity, RBAC decides which servers and tools each may call, and every tool call is logged. Protected databases can be exposed to MCP clients so an assistant can run queries without holding a database password. An MCP Catalog helps teams find and publish secured MCP servers.
Agent Behaviour And Audit
Beams Session Summaries condense an agent’s identity, privileges, tool and API calls, prompts, responses and reasoning into a short readable record. Agentic Classifiers check behaviour against organisation-specific criteria and flag activity that does not fit the agent’s declared goal, including emergent coordination and goal hijacking. Every model download, load or export is also logged.
Availability
Secure MCP access is part of the generally available platform. Beams Session Summaries, Agentic Classifiers and Risk Scoring were previewed at Black Hat USA 2026 through a technology preview programme. Beams itself opens on 9 November 2026, with a waitlist.
Integrations
| Area | Integrations |
|---|---|
| Identity providers | Okta, Microsoft Entra ID, AWS IAM Identity Center, Google Workspace, Keycloak, Active Directory, GitHub SSO |
| Access request workflows | Slack, Microsoft Teams, Jira, PagerDuty, ServiceNow, Discord, Mattermost |
| CI/CD and automation | Jenkins, Argo CD, BuildKite, Harness, TeamCity, Travis CI, Ansible |
| Observability and SIEM | Splunk, Datadog, Elastic, Grafana, Prometheus, InfluxData, Fluentd |
Company Background
| Founded | 2015, as an open-source project |
|---|---|
| Headquarters | Oakland, California, United States |
| Leadership | Ev Kontsevoy (Co-founder and CEO), Alexander Klizhentas (Co-founder and CTO), Taylor Wakefield (Co-founder and COO) |
| Funding | A $110 million Series C led by Bessemer Venture Partners with Insight Partners, Kleiner Perkins and S28 Capital (May 2022) |
| Cisco | A strategic partnership announced in August 2026 covering technology integration, licensing and investment, with Cisco becoming a strategic investor |
| Recognition | Three Cyber Defense Magazine InfoSec Awards (2026); included in the Inc. 5000 list (2026) |
Note: This overview is compiled from Teleport’s publicly available information as of 3 November 2026. It is provided for information only. NHI Mgmt Group does not endorse any vendor, and nothing on this page represents the opinions of NHI Mgmt Group.