Top Google Cloud Agent Identity Alternatives

Auth0 adopts a contemporary method for managing identity, allowing organizations to ensure secure access to applications for all users. It offers a high degree of customization while remaining both straightforward and adaptable. Handling billions of login transactions every month, Auth0 prioritizes convenience, privacy, and security, enabling customers to concentrate on their innovative efforts. Furthermore, Auth0 facilitates quick integration of authentication and authorization processes across web, mobile, and legacy applications, featuring advanced Fine Grained Authorization (FGA) that expands the capabilities of traditional role-based access control, thereby enhancing security measures overall.

Amazon Bedrock AgentCore Identity is a cloud service tailored for managing identities and credentials specifically for AI agents and automated processes. This innovative service empowers developers to create workload identities, configure OAuth 2.0 credential providers, and manage secure tokens that facilitate both user-delegated and machine-to-machine interactions with AWS resources and external services. By effectively managing credentials such as OAuth tokens, it incorporates robust authorization mechanisms to ensure that agents can access necessary resources, either on behalf of users or through their own assigned permissions. The service includes functionalities like inbound authentication methods, secure credential storage, workload identity controls, and smooth integration with both Amazon Bedrock AgentCore Runtime and Gateway. It is primarily aimed at developers and security teams dedicated to building agents that necessitate secure access to a variety of tools, APIs, and data requiring user authorization. Furthermore, AgentCore Identity is compatible with OAuth providers, making it suitable for agent applications operating both within the AgentCore ecosystem and in external environments. This adaptability not only enhances the security of identity management but also fosters modern development practices, ensuring that developers can efficiently navigate the complexities of credential management. As a result, it proves to be an indispensable resource for those working in the field of AI and automation.

Ping Identity delivers a global identity security solution through its advanced identity platform. With a wide array of features including single sign-on (SSO), multifactor authentication (MFA), directory services, and more, it supports enterprises in achieving a delicate balance between security measures and user experience for various identity types, such as workforce, customer, and partner. The platform accommodates different cloud deployment models, such as identity-as-a-service (IDaaS) and containerized software, making it suitable for diverse organizational needs. Ping's solutions are tailored for both developers and IT teams, facilitating seamless digital collaboration via straightforward integrations with widely used tools. These integrations empower employees to work efficiently from any location while utilizing these tools effectively. Furthermore, the platform ensures rapid deployment and interoperability across the entire identity ecosystem. Users can opt for a straightforward SSO solution or implement a more sophisticated, risk-based adaptive authentication system. Additionally, the PingOne package is designed to provide cost efficiency by allowing businesses to pay only for the features they require, all while offering scalability for future growth. This flexibility makes Ping Identity an appealing choice for organizations looking to enhance their identity security framework.

Hush Security provides a robust platform aimed at the recognition and oversight of AI agents alongside other non-human identities. This innovative solution proficiently outlines agent identities, their credentials, permissions, tools, and links to infrastructure, empowering security and identity teams to thoroughly assess risks and implement essential policy measures. Targeted at businesses employing AI agents, the offering features functionalities such as agent inventories, ownership attribution, activity auditing, access governance, and policy enforcement. Hush Security guarantees that organizations retain control over agent permissions, reduce reliance on static credentials, and achieve visibility across both agent and developer environments, including platforms like Claude Code, Cursor, Codex, and Microsoft Foundry. Moreover, teams can utilize the platform to identify agents and credentials, evaluate access levels, set scoped permissions, and monitor agent activities effectively. The free version caters to organizations with up to 25 protected agents, 50 access policies, and 10,000 discovered identities, while its premium plans are tailored for broader applications, making it a flexible solution for diverse organizational requirements. This ability to scale ensures that as organizations expand, their security and governance measures can adapt to evolving needs, thereby enhancing their overall operational resilience. Ultimately, Hush Security positions itself as a vital tool for organizations looking to navigate the complexities of AI agent management and security.

NewCore is an innovative identity provider that caters to both individuals and AI agents through a centralized identity core, which guarantees that every identity remains visible, secure, and continuously monitored from access initiation to auditing completion. It identifies a variety of primary identities, including employees, contractors, partners, machines, delegated agents, and autonomous workloads, all within a cohesive framework. The Identity Security feature reinforces trust in access management with capabilities like single sign-on, passkeys, phishing-resistant multi-factor authentication, secure token signing, strength assessments for authenticators, and security-aware policies. By implementing a Secure Split Key mechanism, it separates signing authority between NewCore’s cloud services and the client's environment, ensuring that neither party can independently validate a token. Additionally, the Identity Discovery feature generates a detailed real-time graph that includes humans, agents, accounts, applications, authenticators, sessions, and access routes, effectively highlighting shadow accounts, orphaned credentials, and unmanaged agents that may introduce security vulnerabilities. This comprehensive approach not only boosts security measures but also simplifies the tracking and management of a wide array of identities within an organization, ultimately fostering a safer digital environment. By integrating advanced technology with robust security protocols, NewCore positions itself as a leader in identity management solutions.

Keycard serves as a sophisticated identity and access management solution designed for the age of agent-driven technologies, ensuring secure interactions between AI agents, users, services, and APIs through real-time identity controls based on established policies. Rather than depending on fixed secrets, the platform produces dynamic access tokens that are short-lived, while also supporting federated identity systems to merge users, agents, and workloads within a decentralized authorization framework. Developers can utilize user-friendly SDKs compatible with major frameworks, allowing them to build applications that recognize agents without requiring deep expertise in identity and access management. The data architecture of the platform includes identity-validated agents, tasks, tools, and resources, which enable the creation of logical zones with context-aware permissions that can be audited. Moreover, security teams have the ability to design deterministic, task-specific policies that define who—whether user or agent—has the authority to execute particular tasks on certain resources under specific conditions, promoting full transparency in access management. This all-encompassing strategy not only bolsters security measures but also streamlines operational processes across diverse systems, ultimately fostering a more integrated technological environment. Through its innovative approach, Keycard paves the way for a future where secure access is paramount for both human and AI interactions.

Hawcx is a cutting-edge cloud-based authentication platform that empowers organizations to bolster security for AI agents while providing passwordless sign-in capabilities for both employees and customers. This innovative solution uniquely links AI agents to specific individuals, guaranteeing that every interaction with the tools adheres to defined policies, while also allowing for the routing of sensitive tasks requiring human oversight, recording decisions made, and facilitating the identification and cancellation of agents when needed. For applications serving both the workforce and consumers, Hawcx offers device-bound, phishing-resistant authentication that incorporates verification processes along with additional security protocols tailored to the assessed risks. Development teams can effortlessly embed passwordless authentication into their applications using web, mobile, and server SDKs, and enterprise logins can be aligned with existing identity management frameworks. Addressing a wide array of needs, the platform is indispensable for security and risk management experts, developers focused on platforms and AI, as well as organizations eager to deploy agents or build web and mobile applications that demand thorough access control and meticulous action tracking. Furthermore, by addressing the increasing necessity for strong security measures in digital spaces, Hawcx positions itself as a vital tool in the evolving landscape of technology.

MudraID is a cloud-centric solution crafted for managing identity and trust, specifically aimed at groups involved in the creation of AI agents, APIs, and services that utilize agent functionalities. The platform endows agents with distinct cryptographic identities, supports scoped and temporary delegation, and evaluates requests in real-time based on established policy, trust, and revocation status. Teams possess the capability to issue, rotate, suspend, and cancel agents' credentials, enforce permissions for each action, and efficiently monitor delegation paths. It aligns with protocols such as OAuth 2.0, OIDC, JWT/JWKS, MCP authorization, and signed A2A Agent Cards, ensuring a strong framework for secure communications. Developers benefit from a Python agent SDK and middleware options designed for FastAPI/Starlette, while an intricately hash-chained audit trail thoroughly documents decisions and related events. MudraID proves to be an indispensable resource for teams in product, platform, security, and IAM domains that need to effectively manage how agents interact with services, thus bolstering overall operational integrity and security. This all-encompassing strategy guarantees that each interaction is not only traceable but also adheres to predefined policies, contributing to a well-governed ecosystem. Moreover, the platform's versatility allows organizations to customize their identity management processes according to their specific operational needs.

The Microsoft Entra Agent ID is designed to streamline identity management and access control for AI agents within a company. IT and identity management teams can create detailed blueprints for agent identities, assign unique identities to individual agents, determine ownership and sponsorship, and manage the lifecycle and permissions of these agents. This platform is available to Microsoft Entra customers and supports OAuth 2.0, enabling communication between agents while also integrating with Microsoft Graph APIs for efficient programmatic management and provisioning. Organizations can connect agents created on various platforms, including Microsoft and external services such as AWS Bedrock and n8n. To leverage enhanced security features like Conditional Access, governance, and risk management for these agents, companies must acquire a Microsoft Agent 365 license, which comes included with the Microsoft 365 E7 package. This functionality is especially advantageous for organizations developing or implementing AI agents that necessitate strong identity, access, and governance systems. By utilizing Microsoft Entra Agent ID, organizations can significantly boost their security measures while adeptly managing their AI agent ecosystems. Moreover, this capability not only simplifies administration but also ensures compliance with evolving security standards.

Permiso is an enterprise identity security platform built to protect every identity operating across cloud, SaaS, on-premises, hybrid, and AI-driven environments. The platform combines identity discovery, runtime identity attribution, identity security posture management, identity threat detection and response, and AI identity security into a unified security architecture. Its Universal Identity Graph creates continuous relationships between users, service accounts, non-human identities, AI agents, credentials, workloads, machines, permissions, infrastructure resources, and runtime actions to provide complete identity lineage. Unlike traditional identity providers that primarily monitor authentication, Permiso extends visibility into runtime activity, allowing security teams to observe tool calls, MCP invocations, serverless functions, workload creation, sub-agent execution, API interactions, and privilege inheritance throughout an identity's lifecycle. The platform inventories identities across cloud providers, SaaS applications, CI/CD pipelines, and enterprise infrastructure while identifying overprivileged accounts, stale credentials, toxic permission combinations, and identities most likely to be compromised. Permiso continuously monitors runtime behavior to detect anomalous activity, lateral movement, credential theft, insider threats, account takeover, and attacks targeting AI agents or machine identities. Security teams can investigate incidents using correlated runtime and control plane activity while leveraging more than 1,500 threat detection signals developed by Permiso's P0 Labs research team. The platform includes dedicated capabilities for identity visibility, identity intelligence, identity posture management, identity threat detection and response, non-human identity security, and AI agent runtime security.

Arcade.dev is an innovative platform tailored for the execution of AI tool calls, enabling AI agents to perform real-world tasks like sending emails, messaging, updating systems, or triggering workflows via user-authorized integrations. Acting as a secure authenticated proxy that adheres to the OpenAI API specifications, Arcade.dev facilitates models' access to a variety of external services such as Gmail, Slack, GitHub, Salesforce, and Notion, utilizing both ready-made connectors and customizable tool SDKs while proficiently managing authentication, token handling, and security protocols. Developers benefit from a user-friendly client interface—arcadepy for Python or arcadejs for JavaScript—that streamlines the processes of executing tools and granting authorizations, effectively removing the burden of managing credentials or API intricacies from application logic. The platform boasts impressive versatility, enabling secure deployments across cloud environments, private VPCs, or local setups, and includes a comprehensive control plane for managing tools, users, permissions, and observability. This extensive management framework guarantees that developers can maintain oversight and control, harnessing AI's capabilities to automate a wide range of tasks efficiently while ensuring user safety and compliance throughout the process. Additionally, the focus on user authorization helps foster trust, making it easier to adopt and integrate AI solutions into existing workflows.

Transform the outdated and vulnerable practices of handling non-human identities by adopting our automated and transparent Workload IAM solution. Manage access between workloads with the same level of oversight you would apply to users, leveraging automated, policy-driven, and identity-focused controls that enable proactive risk mitigation associated with non-human entities. Aembit significantly bolsters security by cryptographically authenticating workload identities in real-time, ensuring that only authorized workloads have access to your sensitive data. By integrating short-lived credentials into requests as necessary, Aembit removes the need for storing or protecting secrets. Moreover, it enforces access permissions dynamically, based on real-time evaluations of workload security posture, geographic location, and other vital behavioral metrics. Aembit proficiently secures access across various environments, including cloud platforms, on-premises infrastructure, and Software as a Service (SaaS) applications, offering a holistic identity management solution. This cutting-edge strategy not only streamlines security processes but also strengthens confidence in your digital infrastructure, ultimately fostering a more resilient environment for your operations.

A production agent often needs the same access as the person who asked for the work. Scalekit supplies that access. It is the gateway those agents call, over MCP, when they have to act inside a customer's own software. The Free plan costs $0 and covers 10,000 tool calls plus 10,000 trigger events each month. Growth costs $99 a month for 100,000 tool calls, with extra calls at $0.50 per thousand. Enterprise is quoted. On that plan the Scalekit dashboard can use SSO and SCIM, and a HIPAA BAA is available. Hosted Scalekit runs in Los Angeles or in Frankfurt. Adding EU residency to Growth costs $99 a month. Larger deployments can sit in the customer's VPC, or on infrastructure the customer runs, with air-gapped installs supported. The trust center lists SOC 2 Type II and ISO 27001. Bartłomiej Krupa, a developer at Brand24, said: "We wanted our users to be able to ask their AI tools the same questions they'd ask Brand24 directly. Scalekit let us ship that without turning our engineering team into OAuth specialists." Under the hood, a request for Salesforce, Gmail, Slack, Notion, HubSpot, Jira, GitHub, or another of the 600+ connectors is matched to that person's stored login, allowed or refused, and sent. The catalog covers 35,000+ tools. Logins are encrypted with AES-256 and kept per tenant. The agent receives a result, not the secret. The same path works for code written with Anthropic, OpenAI, LangChain, Google ADK, Vercel AI, or Mastra.

StackOne is an AI agent integration and execution platform that helps developers connect autonomous agents, AI assistants, and enterprise applications through a unified infrastructure layer. The software provides access to more than 570 prebuilt connectors and 35,000 actions spanning human resources, recruiting, CRM, finance, identity management, customer support, document management, and other business systems. Its integration interfaces support Model Context Protocol, Agent-to-Agent communication, REST APIs, and development SDKs for Python and TypeScript. StackOne enables AI agents to retrieve information, update records, trigger workflows, create objects, and perform authorized operations within connected applications. Managed authentication handles OAuth flows, SAML authentication, API credentials, token refresh, and account connections without requiring developers to implement separate authentication systems for every integration. The AI Connector Builder allows users to develop custom connectors and extend existing integrations using AI-assisted development tools and automated API testing. Its Falcon execution engine processes agent actions while managing retries, API errors, data transformations, authentication, and provider-specific technical requirements. Security capabilities include prompt injection detection, granular action and field permissions, user-scoped access controls, and configurable policies governing agent behavior. Observability tools provide request tracing, audit logs, and operational monitoring to help administrators understand which actions agents perform and which resources they access. Dynamic tool discovery, managed data synchronization, and token optimization help reduce context window usage while improving the relevance of tools available during agent execution.

Idira is an innovative identity security platform developed by Palo Alto Networks, specifically designed for enterprises leveraging artificial intelligence, and it focuses on protecting various forms of identities—whether human, machine, or agentic—through a unified control system. This platform transforms the landscape of privileged access management by expanding the oversight of privileges to include all identities capable of engaging with sensitive systems, which involves a wide array of elements such as data, applications, cloud services, workloads, endpoints, secrets, certificates, SSH keys, and AI agents. By pinpointing potential identity risks, Idira dynamically modifies access privileges and oversees the entire lifecycle from the moment access is granted to the end of a session. Transitioning away from a conventional, always-available access model, it prioritizes dynamic privileges, just-in-time access, and zero standing privilege while ensuring ongoing verification and policy-driven controls with real-time enforcement that considers identity, device, and contextual factors. Furthermore, for human identities, Idira integrates privileged access management, workforce access, endpoint privilege management, and identity governance, equipping organizations with essential tools to effectively combat privilege sprawl. This holistic strategy not only fortifies security measures but also enhances operational efficiency across the entire organization, ultimately leading to a more secure and streamlined environment for all users.

Defakto Security presents a powerful platform that authenticates all automated interactions by issuing temporary, verifiable identities to non-human entities such as services, pipelines, AI agents, and machines, effectively eliminating the reliance on static credentials, API keys, and persistent privileges. Their extensive non-human identity and access management solution supports the detection of unmanaged identities across various environments, including cloud, on-premises, and hybrid configurations, allowing for the real-time issuance of dynamic identities in accordance with policy requirements, the enforcement of least-privilege access principles, and the creation of comprehensive audit-ready logs. The solution consists of multiple modules: Ledger, which guarantees continuous discovery and governance of non-human identities; Mint, which streamlines the generation of targeted, temporary identities; Ship, which supports secretless CI/CD workflows by removing hard-coded credentials; Trim, which refines access rights and removes excessive privileges for service accounts; and Mind, which protects AI agents and large language models using the same identity framework utilized for workloads. Each module is essential in bolstering security and optimizing identity management across a variety of operational landscapes. Together, these components not only enhance security but also promote efficiency in managing identities for non-human entities.

SystemPrompt serves as a comprehensive platform aimed at organizations, offering self-hosted AI governance and runtime capabilities that connect AI clients, model providers, and an array of tools. It features a centralized control plane and gateway for managing user and agent access, directing model routing, and ensuring the enforcement of permissions, credentials, and compliance standards. In addition, the platform facilitates the documentation of requests and tool interactions within the organization, aiding in auditing and review processes. With support for connections to providers like Anthropic, OpenAI, Google Gemini, and Groq, SystemPrompt utilizes the Model Context Protocol (MCP) to guarantee that only authorized tools are made available to the right clients. Moreover, the platform enhances operational efficiency by allowing teams to integrate agents, workflows, skills, and custom business services, all of which can be deployed through binaries, containers, or managed cloud environments. Tailored specifically for technical and security teams overseeing AI workloads, SystemPrompt provides local evaluation at no charge while requiring that production use and cloud hosting comply with commercial agreements. Additionally, its adaptable architecture empowers organizations to customize the platform according to their unique operational requirements, ensuring greater flexibility and responsiveness to changing needs.

Peta acts as a sophisticated control plane for the Model Context Protocol (MCP), facilitating, securing, regulating, and supervising the interactions between AI clients and agents with external resources, data, and APIs. The platform incorporates a zero-trust MCP gateway, a secure vault, a managed runtime environment, a policy engine, human-in-the-loop approvals, and extensive audit logging into a unified solution, allowing organizations to enforce detailed access controls, protect sensitive credentials, and track all interactions performed by AI systems. Central to Peta is Peta Core, which serves as both a secure vault and gateway, responsible for encrypting credentials, generating ephemeral service tokens, ensuring identity verification and policy compliance for each request, managing the lifecycle of the MCP server through lazy loading and auto-recovery, and injecting credentials at runtime without exposing them to agents. Furthermore, the Peta Console enables teams to determine which users or agents can access specific MCP tools within defined environments, set up approval processes, manage tokens, and analyze usage data along with associated costs. This comprehensive strategy not only bolsters security but also promotes effective resource management and accountability across AI operations, ultimately leading to improved operational efficiency and enhanced oversight. By integrating these functionalities, Peta establishes a robust foundation for organizations seeking to optimize their AI-driven initiatives.

Smallstep delivers the first true Device Identity Platform™, enabling enterprises to enforce Zero Trust at the device level. By leveraging ACME Device Attestation, it binds identity to hardware co-processors, ensuring credentials cannot be stolen, copied, or reused on unauthorized machines. This allows organizations to confidently secure access to Wi-Fi networks, VPNs, SaaS applications, cloud APIs, Kubernetes, Git repositories, and regulated data. Unlike traditional certificate or MDM-based approaches, Smallstep provides cryptographic proof that a device is both authentic and company-owned. The platform supports all major operating systems and integrates with over 100 enterprise tools, making deployment seamless at scale. Trusted by Fortune 100 companies, Smallstep helps IT, security, and DevOps teams eliminate device identity gaps. The result is stronger security, reduced attack surface, and Zero Trust that actually works in modern environments.

Eyrane represents a cutting-edge AI security platform currently under development, tailored for teams engaged in security and engineering, particularly those overseeing agents and automated processes. Its core purpose is to aid organizations in pinpointing an array of elements such as agents, models, tools, MCP servers, identities, credentials, permissions, and interconnected systems, while also charting the access routes between these various components. The platform is designed to implement runtime policies that determine whether actions by agents are authorized, need further scrutiny, or are outright forbidden, based on criteria such as identity, permissions, tools, and the sensitivity of the data at play. Furthermore, Eyrane assesses the performance of agents in the context of potential threats, including prompt injection, tool misuse, privilege abuse, data exfiltration, and issues related to MCP. Its potential uses include managing agent inventories, conducting access audits, real-time monitoring, bolstering MCP security, and supporting AI red teaming efforts. Eyrane ultimately strives to empower security and engineering teams in deploying AI agents across business applications and workflows, ensuring thorough oversight and control throughout the process. With its emphasis on balancing security and operational efficiency, Eyrane emerges as an essential asset in the dynamic realm of AI governance, poised to adapt to future challenges as they arise.

Bevel functions as a vendor-agnostic control plane integrated with Git, specifically designed for enterprise AI agents, enabling organizations to establish their agents, context, skills, tools, permissions, and identities as proprietary files within their systems, accessible by any agent runtime through the Managed Control Plane (MCP). The context is structured as categorized knowledge nodes, complete with documented provenance that outlines the source, last modification, and verification timestamps, all of which is synthesized into a navigable graph that can be continuously updated for dashboard creation. Skills are defined in clear Markdown formats, allowing process owners to effortlessly read and review modifications, as well as transfer them between various runtimes. Tool manifests detail the available functionalities, while sensitive data is securely encrypted in a vault, managed by access protocols that specify which agents have permission to read particular files or execute certain endpoints. Each agent is allocated a distinctive identity and set of credentials, ensuring traceability of all actions to their origin. This extensive framework not only fortifies security and structure but also fosters transparency and accountability within AI operations, creating a more robust ecosystem for enterprise-level management of AI agents. Furthermore, the system promotes collaborative development, allowing teams to innovate while maintaining control over their AI resources.

Token Security introduces a groundbreaking strategy designed specifically for the rapidly growing domain of Non-Human Identities (NHI), advocating for a machine-centric method to identity protection. In this digital age, identities are everywhere and frequently remain unmonitored; they emerge from machines, applications, services, and workloads that are created by diverse sources throughout each day. The complex and sluggish process of overseeing these identities has expanded the attack surface, making it challenging for organizations to manage effectively. Instead of focusing exclusively on human identities, Token emphasizes the significance of the resources being accessed, promptly illuminating who interacts with which resources, pinpointing vulnerabilities, and ensuring robust security without hampering operations. Additionally, Token proficiently maps all identities within cloud ecosystems, seamlessly incorporating complex elements such as Kubernetes, databases, servers, and containers, which leads to a unified view of critical identity data. This all-encompassing methodology not only bolsters security but also streamlines identity management amidst increasingly intricate infrastructures, ultimately fostering a more resilient digital environment. As organizations increasingly rely on automation and interconnected systems, the need for such innovative identity solutions becomes even more crucial, showcasing Token's relevance in today's technological landscape.

Composio is a powerful platform that transforms AI agents from simple conversational tools into fully functional systems capable of executing complex workflows. It provides seamless integration with over 1,000 applications, including popular tools like Slack, Gmail, GitHub, Notion, and more. The platform handles authentication automatically, managing OAuth flows, API keys, and permissions without requiring manual setup. Composio enables agents to dynamically select and execute the right tools based on user intent, improving efficiency and accuracy. It also offers secure, sandboxed environments where workflows can run safely, ensuring isolation and reliability. Developers can build multi-step workflows, automate processes, and execute tasks in parallel for greater productivity. The platform supports context-aware sessions, allowing agents to retain state and continue tasks without restarting. Composio is model-agnostic, giving developers the freedom to switch between AI models without disrupting their workflows. Its infrastructure is designed for scalability, supporting enterprise-grade use cases and high-volume operations. With built-in security features and compliance standards, it ensures safe handling of data and operations. By simplifying integrations and execution, Composio reduces development overhead and accelerates deployment. Ultimately, Composio enables developers to build intelligent, action-oriented AI agents that can operate across the entire software ecosystem.

Most companies write permission checks straight into each application, which makes access hard to change, hard to review and hard to prove to an auditor. Cerbos moves those rules into one authorization management platform that covers applications, API gateways, service meshes, data platforms and AI agents. Teams describe who can do what in short policy files that people outside engineering can read. Role, attribute and relationship rules (RBAC, ABAC, ReBAC and PBAC) live side by side. When a request comes in, Cerbos pulls the current roles and attributes from Okta, Microsoft Entra ID, your databases or internal APIs, then checks the request against the policy. A change to the rules ships after its tests pass, with no application release. What each team gets For engineering, one API call takes the place of permission logic spread across services, with SDKs in eight languages including Go, Java, Python and JavaScript. For identity teams, a single place to manage access for employees, customers, services and AI agents, next to the identity provider they already run. For product teams, custom roles and per-tenant permissions for enterprise customers without building them one customer at a time. For security teams, every AI agent tool call goes through the same policy check as a person, so agent rollouts can be approved with controls in place. Each check is recorded along with the policy that applied, which gives SOC 2, ISO 27001 and HIPAA audits a clear trail. Deploy it as a hosted service, in your own cloud or fully air-gapped. Customers running Cerbos in production include Utility Warehouse, 9fin and Salesroom.

Google has launched the Agent Payments Protocol (AP2), a collaborative and open protocol co-developed with over 60 varied companies in the realms of payments, fintech, and technology, including major players like Mastercard, PayPal, Adyen, Coinbase, and Etsy, with the purpose of enabling secure transactions conducted by agents across multiple platforms. This innovative protocol expands upon earlier open standards such as Agent2Agent (A2A) and the Model Context Protocol (MCP), ensuring that when an AI agent handles a payment on behalf of a user, it meets three critical standards: authorization, confirming the user's explicit consent for the transaction; authenticity, ensuring that the agent's intended purchase matches the user's true intent; and accountability, which preserves clear audit trails and assigns responsibility for any mistakes or fraudulent activities. To maintain these rigorous standards, the protocol integrates mandates, which are cryptographically signed digital agreements backed by verifiable credentials, thus bolstering security and fostering trust in agent-driven transactions. By introducing AP2, Google aims to make a notable leap forward in the digital payments landscape, striving to boost users' confidence in their automated financial dealings. This initiative not only enhances transaction security but also positions Google as a leader in redefining how digital payments are processed in an increasingly automated world.

Descope serves as a CIAM platform designed specifically for developers and IT teams. With our intuitive no-code workflow builder and SDKs, customers can effortlessly design seamless and secure user journey flows for every interaction. By taking care of authentication methods, session management, and risk management, Descope empowers developers to focus more on their primary app projects. The platform simplifies the management of users, tenants, RBAC, SCIM provisioning, and identity federation, ensuring a hassle-free experience. Through Descope, customers can enhance user experience with passwordless authentication, bolster security with risk-based MFA, and unify user identities across all applications used in business, ultimately creating a more streamlined operational environment. This comprehensive approach not only improves security but also fosters greater user satisfaction and trust.

Totem is a cloud-centric framework tailored for AI agents that enables organizations to recognize, connect, and manage agents across different boundaries. Each agent possesses a verified identity, which allows for controlled access to organizational communications, facilitates direct messaging among agents, supports collaborative task management, and streamlines various workflows. Teams are empowered to register agents, manage their credentials and permissions, and can opt for either a web dashboard or a command-line interface for their connection settings. Developers can also take advantage of an HTTPS API to effortlessly integrate Totem into their existing applications and automation workflows. The platform accommodates a wide range of applications, such as procurement, data access governance, customer support transitions, vendor management, and compliance interactions. Furthermore, Totem guarantees the validation of an agent’s identity, their interconnections, and the permissions related to their actions, ensuring that it does not engage in making business decisions for users. This clear delineation of roles not only boosts operational efficiency but also fortifies the security of organizational processes involving AI agents, ultimately leading to a more streamlined approach to managing complex interactions.

NVIDIA OpenShell is an open and secure runtime environment tailored for the operation of autonomous AI agents, managing their execution, access privileges, and the routes for inference traffic. Rather than incorporating security measures directly into the model or application, it upholds safety protocols in the surrounding infrastructure, ensuring that nothing is permitted by default; permissions are allocated through predefined policies, with enforcement mechanisms acting outside the agent's process to avert any potential bypass through prompts. Each autonomous agent is confined within a controlled sandbox, which limits direct network access, restricts file visibility, and employs kernel-level monitoring of system calls to guarantee stringent oversight. Serving as the control plane, a gateway handles user authentication, manages the lifecycle of sandboxes, and provides the agents with policies, configurations, credentials, and settings for inference. Furthermore, an external supervisor evaluates network requests based on policies at multiple levels—including binary, destination, method, and path—and grants access to credentials only when expressly authorized, thereby bolstering the overall security architecture. This comprehensive and layered framework ensures that the agents operate efficiently while upholding strict security measures throughout their processes, ultimately fostering a trustworthy environment for AI interaction. The attention to detail in security design reflects a commitment to safeguarding both the agents and the systems they interact with.

Logto serves as a contemporary alternative to Auth0, tailored for software as a service (SaaS) and application developers, making it an excellent option for both expanding businesses and individual users. It provides a comprehensive identity management solution with SDKs that facilitate seamless authentication processes. Users can choose from a variety of sign-in options, including social logins and passwordless methods, while having the ability to personalize UI components to align with their branding. The platform is built with a ready-to-use infrastructure, eliminating the necessity for additional setup. It also includes a management API that is immediately accessible, along with flexible connectors designed for customization and scalability, supporting SAML, OAuth, and OIDC protocols. Logto is equipped for enterprise needs, featuring role-based access control (RBAC), organizational capabilities for multi-tenant applications, user management tools, audit logs, single sign-on (SSO), and multi-factor authentication (MFA), ensuring strong security and compliance for its users. Additionally, its user-friendly interface and extensive support resources make it an ideal choice for developers seeking to enhance their application's authentication experience.

The UCP and AP2 documentation details the seamless integration of the Universal Commerce Protocol (UCP) with the Agent Payments Protocol (AP2), facilitating secure and verifiable transactions executed by AI agents or platforms on behalf of users, thus enabling commerce systems to handle discovery, checkout, and payment processes independently of intermediaries. The full compatibility between UCP and AP2 creates a trust framework for agent-led transactions, which requires a secure and cryptographically verifiable exchange of intent and authorization between businesses and platforms through Verifiable Digital Credentials (VDCs); this system ensures that businesses obtain signed checkout commitments that remain unchanged throughout the transaction process, while platforms deliver proofs of payment authorization that are directly associated with the state of the cart, significantly reducing the risk of fraud and confirming that transactions are both conclusive and authentic. Moreover, this integration not only streamlines the transaction process but also significantly boosts the overall efficiency and reliability of digital commerce ecosystems, ultimately fostering greater confidence among all participants in the marketplace.