# Identra

> Identra (MeruMesh Inc) is an identity security platform: one control plane that secures human identities, non-human identities (service accounts, workloads, OAuth apps, tokens, keys), and AI-agent identities, from discovery to detection and response, at runtime. Identra Guard extends this into the browser, protecting identities, AI usage, and data inside the live session.

## Company

- [Homepage](https://www.identra.ai/): what Identra secures and how the platform works
- [Blog](https://www.identra.ai/blog/): Identra's research and points of view on identity security

## Key content

- [The Non-Human Majority](https://www.identra.ai/blog/non-human-majority/): Identra's flagship essay on non-human identity. Most enterprise identities are not human; securing service accounts, tokens, OAuth apps, workloads, and AI agents is the defining security problem of the agentic era. Covers the accountability gap, the three asymmetries of machine identity, the four questions every identity must answer, authority fan-out from AI agents, and the control loop for governing non-human identity.
- [The Non-Human Identity Assessment](https://www.identra.ai/blog/non-human-majority/assessment/): an interactive maturity matrix and scorecard for rating a non-human identity security program across six dimensions.
- [Keys That Never Die](https://www.identra.ai/blog/keys-that-never-die/): Identra Research audit of expiry defaults for 20 credential configurations across 19 major platforms. Only 7 configurations expire by default, 5 are documented to never expire, 8 have no documented default expiration, and only 9 come with a documented admin control that enforces a maximum lifetime. Includes the full dataset and methodology, checked against vendor documentation on July 12, 2026.
- [Securing AI Agent Identities: A Practical Guide](https://www.identra.ai/blog/securing-ai-agent-identities/): Identra's practical guide to securing AI agent identities. Covers directory-native agent identity with named owners and identity levels from blueprint to execution, least-privilege short-lived credentials, deterministic controls layered with behavioral detection, consequence-based approval gates, multi-layer kill switches and safe action logging, delegation profiles built on RFC 8693 token exchange, MCP proxy confused-deputy controls, fleet governance, and a 30-60-90 day rollout plan.

## Glossary

- [Identity Security Glossary](https://www.identra.ai/glossary/): definitional answers on human, non-human, and AI-agent identity security, from secret sprawl to MCP security
- [What is identity security?](https://www.identra.ai/glossary/identity-security/)
- [What is identity threat detection and response (ITDR)?](https://www.identra.ai/glossary/identity-threat-detection-response/)
- [What is identity security posture management (ISPM)?](https://www.identra.ai/glossary/identity-security-posture-management/)
- [What is account takeover (ATO)?](https://www.identra.ai/glossary/account-takeover/)
- [What is session hijacking?](https://www.identra.ai/glossary/session-hijacking/)
- [What is phishing-resistant MFA?](https://www.identra.ai/glossary/phishing-resistant-mfa/)
- [What does zero trust mean for identity?](https://www.identra.ai/glossary/zero-trust-identity/)
- [What is a non-human identity (NHI)?](https://www.identra.ai/glossary/non-human-identity/)
- [What is the difference between a non-human identity and a machine identity?](https://www.identra.ai/glossary/nhi-vs-machine-identity/)
- [What is a machine identity?](https://www.identra.ai/glossary/machine-identity/)
- [What is a service account and why is it risky?](https://www.identra.ai/glossary/service-account-security/)
- [What is workload identity?](https://www.identra.ai/glossary/workload-identity/)
- [What are OAuth app risks?](https://www.identra.ai/glossary/oauth-app-risk/)
- [What is API key security?](https://www.identra.ai/glossary/api-key-security/)
- [What is secrets management?](https://www.identra.ai/glossary/secrets-management/)
- [What is secret sprawl?](https://www.identra.ai/glossary/secret-sprawl/)
- [What is non-human identity lifecycle management?](https://www.identra.ai/glossary/nhi-lifecycle/)
- [What is identity sprawl?](https://www.identra.ai/glossary/identity-sprawl/)
- [What is just-in-time (JIT) access?](https://www.identra.ai/glossary/just-in-time-access/)
- [What is zero standing privileges (ZSP)?](https://www.identra.ai/glossary/zero-standing-privileges/)
- [What is the principle of least privilege?](https://www.identra.ai/glossary/least-privilege/)
- [What is the difference between PAM and IAM?](https://www.identra.ai/glossary/pam-vs-iam/)
- [What is the OWASP Non-Human Identity Top 10?](https://www.identra.ai/glossary/owasp-nhi-top-10/)
- [What is an AI agent identity?](https://www.identra.ai/glossary/ai-agent-identity/)
- [What are the security risks of MCP (Model Context Protocol)?](https://www.identra.ai/glossary/mcp-security/)
- [What is agentic AI security?](https://www.identra.ai/glossary/agentic-ai-security/)
- [What is shadow AI?](https://www.identra.ai/glossary/shadow-ai/)
- [What is enterprise AI governance?](https://www.identra.ai/glossary/ai-governance/)
- [What is prompt injection?](https://www.identra.ai/glossary/prompt-injection/)
- [What is the confused deputy problem?](https://www.identra.ai/glossary/confused-deputy-problem/)
- [What is AI agent delegation (on-behalf-of)?](https://www.identra.ai/glossary/ai-agent-delegation/)
- [What is AI agent sprawl?](https://www.identra.ai/glossary/ai-agent-sprawl/)
- [What is enterprise browser security?](https://www.identra.ai/glossary/enterprise-browser-security/)
- [What is AI data leakage?](https://www.identra.ai/glossary/ai-data-leakage/)
- [What is runtime identity security?](https://www.identra.ai/glossary/runtime-identity-security/)
- [What is workload identity federation?](https://www.identra.ai/glossary/workload-identity-federation/)
- [What are ephemeral credentials?](https://www.identra.ai/glossary/ephemeral-credentials/)
- [What is a token vault?](https://www.identra.ai/glossary/token-vault/)
- [What is the OAuth on-behalf-of flow for AI agents?](https://www.identra.ai/glossary/oauth-on-behalf-of/)
- [What is the A2A protocol?](https://www.identra.ai/glossary/a2a-protocol/)
- [What are AI agent guardrails?](https://www.identra.ai/glossary/ai-agent-guardrails/)
- [How does MCP authentication work?](https://www.identra.ai/glossary/mcp-authentication/)
- [What is CIEM?](https://www.identra.ai/glossary/ciem/)
- [What is SPIFFE?](https://www.identra.ai/glossary/spiffe/)
- [What is token theft?](https://www.identra.ai/glossary/token-theft/)
- [What is consent phishing?](https://www.identra.ai/glossary/consent-phishing/)
- [What is excessive agency?](https://www.identra.ai/glossary/excessive-agency/)
- [What is an AI agent registry?](https://www.identra.ai/glossary/agent-registry/)
- [What is an identity fabric?](https://www.identra.ai/glossary/identity-fabric/)
- [What is Microsoft Entra Agent ID?](https://www.identra.ai/glossary/entra-agent-id/)

## Comparisons

- [Why Posture Tools Miss Runtime Identity Risk](https://www.identra.ai/why/posture-tools-miss-runtime-identity-risk/): Identra's pillar essay on why posture tools (ISPM, CSPM, CIEM) score configuration but cannot see what a valid session does, and why stolen tokens, hijacked AI agents, and awakened NHIs pass every posture check
- [ITDR vs ISPM](https://www.identra.ai/compare/itdr-vs-ispm/): ITDR detects and responds to active identity attacks; ISPM finds and fixes identity misconfigurations before an attack
- [ISPM vs CIEM](https://www.identra.ai/compare/ispm-vs-ciem/): CIEM governs cloud infrastructure entitlements in depth; ISPM assesses posture across the whole identity fabric
- [NHI Security vs Secrets Management](https://www.identra.ai/compare/nhi-security-vs-secrets-management/): secrets management protects credentials; NHI security governs the identities behind them across their whole lifecycle
- [Machine Identity vs Workload Identity](https://www.identra.ai/compare/machine-identity-vs-workload-identity/): workload identity is a subset of machine identity; how the terms differ and where AI agents fit
- [Agentic AI Security vs LLM Security](https://www.identra.ai/compare/agentic-ai-security-vs-llm-security/): LLM security protects a model's inputs and outputs; agentic AI security governs autonomous agents that hold identities and take actions
- [Enterprise Browser vs SWG and SASE](https://www.identra.ai/compare/enterprise-browser-vs-swg-sase/): network-layer SWG and SASE filter traffic in transit; enterprise browser security sees identity and data in the live session