PlainID is an authorisation company that controls what people, non-human identities and AI agents can access, do and expose once they are signed in. It is built for large enterprises that manage authentication centrally but still leave access rules hard-coded inside each application, database and API.
PlainID replaces those scattered rules with one central policy layer. Policies are written once and enforced at runtime wherever access happens: in applications, API gateways, data platforms, identity providers, MCP servers and AI agent frameworks.
At A Glance
| Vendor | PlainID |
|---|---|
| Category | Runtime authorisation, Policy-Based Access Control (PBAC) |
| Identities covered | Workforce users, customers, non-human identities, APIs and services, AI agents |
| Deployment | SaaS policy administration, with decision points in PlainID Cloud or in the customer environment |
| Agentic AI | MCP Gateway, AI-Ready Gateway Enforcement, five agent guardrails, Daia assistant |
| Certifications | ISO 27001:2022, AICPA SOC 2 |
| Funding | Over $100 million raised |
| Customers | Cisco, Citibank, Bank Leumi, Samsung, Wells Fargo |
What PlainID Does
Most organisations know who has logged in, but the question of what that identity may do next is answered inside each application, usually through static roles that drift over time. PlainID treats that question as a shared service. Its model is Policy-Based Access Control, a centralised form of attribute-based control that can also express role-based and relationship-based rules.
The platform follows one lifecycle. Discover finds identities, resources, existing policies and the access paths between them. Manage is where teams write, test, approve and certify policy. Enforce hands decisions to the points where access occurs, either by answering live requests from applications and gateways or by translating policy into the native controls of platforms such as Snowflake and Databricks.
Products
PlainID Runtime Authorization Platform
Access rules that live in dozens of separate codebases are hard to see, let alone change. The core platform combines a central Policy Administration Point, Policy Decision Points and Policy Information Points that pull context from authoritative sources. Business and security teams can build policy in a visual console, while developers can manage the same policy as code. Governance features include versioning, approvals, promotion between environments, a policy simulator for impact testing, certification campaigns and Policy360, a connected view of each policy, its dependencies and its change history.
PlainID Authorizers
Authorizers save developers from writing access logic into every API, microservice and application. Authorizers are pre-built enforcement components that sit in front of or inside the systems being protected and call the decision point on every request. They are catalogued in the PlainID Integration Hub, grouped by how each integration takes part in authorisation, and cover API gateways, service mesh, data platforms, identity providers and AI frameworks.
Policy Orchestration
Data and platform teams often maintain permissions by hand in each tool, which makes audits slow. For platforms that have their own access model, Policy Orchestration Points connect to the third-party system, read its objects and existing policies, and keep them governed from PlainID. Discovery can run on demand or on a schedule and picks up new, changed or deleted objects. Identity attribute mapping across vendors is available for Power BI, Snowflake and Databricks.
Data Access Authorisation
Analysts, applications and agents querying shared data often need different views of the same tables. Data access controls apply row filters, column filters and masking to queries and responses, so a user or agent sees only the records and fields its policy allows. A dedicated container connects to the PlainID runtime to apply fine-grained control to SQL queries.
Structured Rego And AuthZEN
Engineering teams that already run policy as code can keep PlainID policy in their own review workflow. Policies can be exported to Structured Rego, a structured form of the Rego language used by Open Policy Agent, and imported back through the Management APIs. Exports keep a stable ordering, so policy changes show up as small, reviewable diffs in Git. The decision point also supports the OpenID AuthZEN Authorization API 1.0, alongside OpenID Connect, RFC 9396, SCIM and NIST SP 800-207.
How It Works
PlainID separates where policy is managed from where decisions are made, so authorisation can run close to the systems it protects.
| Component | What it does |
|---|---|
| Control plane | PlainID Cloud is a SaaS console organised into tenants, environments, scopes and workspaces, with separate workspaces for identities, policies and integrations. |
| Policy Authorization Agent (PAA) | Each tenant has a cloud-hosted decision point. Organisations that need lower latency, or want attribute data to stay on their own network, run a PAA in their cloud, hybrid or on-premises environment, typically one per environment. |
| PAA components | A Hybrid Agent talks to PlainID Cloud, a PIP reads attributes from local databases, REST APIs and user stores, a PDP returns decisions, and Redis handles local caching. Production setups can use a managed Redis service such as AWS ElastiCache or Azure Redis. |
PAAs deploy through a Helm chart, as a Docker container, as a standalone service for VM-based setups, or on Amazon ECS. Container images are signed and can be verified with Cosign.
The Policy Model
Each policy answers three questions. Who is defined by Dynamic Groups, rules over identity attributes such as all agents owned by the finance team. What is defined by rulesets over asset types and their attributes, together with the permitted actions. When is defined by conditions such as IP range, authentication method or a device posture signal. Every policy resolves to Allow or Deny, and decisions can return filters, masks, token claims or obligations such as step-up authentication rather than a simple yes or no.
Identities And Systems Covered
| Workforce users | Employees and contractors, with context-aware policy replacing static roles in the existing identity provider. |
|---|---|
| Customers and partners | One policy model for workforce, consumer and B2B access. |
| Non-human identities | Microservices and service-to-service connections, governed with zero standing privileges. |
| APIs and services | API routes and service traffic, protected by Authorizers deployed near the workload. |
| AI agents | Agents registered as identities with their own attributes, alongside MCP servers, tools and vector databases. |
| Data systems | Databricks, Snowflake, Google BigQuery, Redshift, PostgreSQL, Oracle, MongoDB, MSSQL, Pinecone, Starburst, Trino and Power BI. |
Agentic AI Capabilities
Composed Identity
PlainID links each agent action to the person or system the agent is working for. This Composed Identity lets a single policy evaluate the user and the agent together, covering both user-to-agent and agent-to-agent transactions.
Five Guardrails
| Guardrail | What it controls |
|---|---|
| Prompt | Blocks queries the user and agent are not entitled to make |
| Retrieval | Filters data before retrieval, so unauthorised content never reaches the model in RAG pipelines |
| MCP and tool use | Decides which servers, tools and parameter values an agent may use |
| API actions | Checks each action an agent takes against an API |
| Output | Masks and filters responses to stop sensitive data leaking |
MCP Gateway
The MCP Gateway, part of PlainID Edge, sits between agents and an organisation’s MCP servers and exposes a single endpoint. It discovers servers and tools, classifies each tool by purpose, sensitivity and risk, and enforces policy at server, tool and parameter level on every call, with no change to existing agents.
AI-Ready Gateway Enforcement
This extends enforcement to an organisation’s existing API, AI and MCP gateways. It evaluates agent intent at the moment of action, governs tool parameters, masks data and keeps an auditable record of each agent decision.
Agent Discovery And Daia
PlainID discovers and registers agents from AWS Bedrock AgentCore and Microsoft Foundry using read-only permissions, and offers authorizers for LangChain and LangGraph. Daia is an assistant that guides policy creation for agentic AI, and the PlainID MCP Server lets administrators query their tenant and write policy in natural language within their own permissions.
Availability
AI-Ready Gateway Enforcement was released in October 2026, and the wider set of agentic capabilities is offered through an early access programme. Cisco Duo Inline Hooks, announced with Cisco in July 2026, let Duo call PlainID during sign-in to add policy-based claims to tokens and SAML assertions.
Integrations
| Area | Integrations |
|---|---|
| Identity providers | Active Directory, Auth0, Cisco Duo, ForgeRock, Microsoft Entra ID, Okta, Ping Identity, Thales OneWelcome |
| Identity governance | SailPoint, Saviynt |
| API and AI gateways | Amazon API Gateway, Apigee, Azure API Management, Envoy, Kong, MuleSoft, Spring Cloud Gateway, WSO2 |
| Service mesh | Istio |
| Data security | BigID, Denodo |
| Agentic platforms | Microsoft Agents SDK, OpenClaw |
| Secrets for PlainID connections | HashiCorp, AWS Secrets Manager, Azure Key Vault, CyberArk CCP |
| Other platforms | Salesforce, Zscaler Private Access |
Company Background
| Founded | 2014, with origins in the banking sector. |
|---|---|
| Headquarters | Tel Aviv, Israel and New York, United States. |
| Leadership | Oren Ohayon Harel (CEO and Co-Founder), Gal Helemski (CPO and Co-Founder). |
| Investors | Insight Partners, which led the Series C, Viola Ventures, Capri Ventures, iAngels and INCAPITAL. |
| Standards work | Member of the OpenID Foundation AuthZEN Working Group. |
| Recognition | Included in the KuppingerCole Leadership Compass for Policy Based Access Management (2025), the Gartner Market Guide for Guardian Agents (2026) and the Gartner Hype Cycles for Digital Identity and for APIs (2026). |
Note: This overview is compiled from PlainID’s publicly available information as of 4 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.