--- sourceDocument: Brazil Workflow Data Fabric sourceDocumentLink: https://www.servicenow.com/docs/r/integrate-applications Release : - brazil ft:locale : - en-US ft:publication_title : - Brazil Workflow Data Fabric ft:clusterId : - crint bundleId : - crint workflow : - Creator --- # Model Context Protocol connectors # Model Context Protocol connectors {#ariaid-title1} Release version: Brazil Updated September 10, 2026 ![](https://www.servicenow.com/docs/portal-asset/ico-clock) 2 minutes to read Summarize ![AI sparkle icon](https://servicenow.com/docs/portal-asset/ai-sparkle-icon) Summarized using AI This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete. ## Summary of Model Context Protocol connectors Model Context Protocol (MCP) connectors in Connect Hub enable ServiceNow AI agents to seamlessly access tools and data from external systems through a standardized open communication protocol. MCP defines how AI agents discover and invoke external tools, retrieve resources, and share context, allowing integration with diverse third-party applications without custom coding. Each MCP connector represents a configured link between ServiceNow and an MCP-compatible external system, facilitating context-aware workflows across AI models and external services. Show full answer Show less ## Key Features * **MCP Registry:** A private ServiceNow registry where AI models, agents, and users discover and connect to approved MCP servers. It ensures governance by allowing approval of servers through AICT, aligning autonomous AI decisions with enterprise policies. The registry integrates with AI Agent Studio and AI platforms like Claude. * **OAuth 2.1 Connector Registration Methods:** Connect Hub supports two OAuth client registration methods: * **Dynamic Client Registration:** Automates OAuth application setup by negotiating client credentials during connection, reducing manual steps and improving security via automatic credential rotation. * **Manual Client Registration:** Requires creating and configuring an OAuth app manually on the third-party system, then entering the client ID and secret into Connect Hub. * **Certified Connectors:** MCP connectors tested and approved by ServiceNow are labeled "Certified by ServiceNow" in the registry, offering trusted integration options. * **Custom Connectors:** If no suitable connector exists in the MCP registry, customers can define custom connectors by specifying the MCP endpoint URL and authentication details. * **Authentication Methods for Custom Connectors:** Supports multiple methods to align with external system security requirements: * OAuth 2.1 (recommended), supporting both dynamic and manual client registration * API key authentication using static keys * Basic Authentication using username and password ## What This Enables for ServiceNow Customers By leveraging MCP connectors, ServiceNow customers can integrate AI agents with a broad range of external tools and data sources efficiently and securely. This integration promotes coordinated, context-aware automation and AI-driven workflows across systems without custom integration coding. The use of MCP Registry and certified connectors ensures governance and simplifies discovery of trusted external MCP servers. OAuth 2.1 support with dynamic client registration reduces setup complexity and enhances credential security. Custom connector support and flexible authentication options provide adaptability for unique enterprise environments. Model Context Protocol (MCP) connectors in Connect Hub enable AI agents to access tools and data from external systems using a standardized communication protocol. MCP is an open protocol that defines how AI agents communicate with external systems to discover and invoke tools, retrieve resources, and share context. By standardizing this communication layer, MCP allows AI agents to interact with a wide range of third-party applications without requiring custom integration logic for each one. In Connect Hub, an MCP connector represents a configured connection between ServiceNow and an external system that exposes an MCP-compatible server. Once an MCP connector is set up, AI agents can use it to access the tools and capabilities that the external system provides, enabling coordinated, context-aware workflows across models and systems. ## MCP Registry {#model-context-protocol-connector__section_hwj_sx2_mjc} ServiceNow MCP Registry is a private registry where large language models (LLMs), AI agents, and teams discover and connect to approved MCP servers, finding the best tool for every business problem, governed at every step. The Registry enables AI models and humans query available MCP servers and select the ones best suited to solve a business problem. Every server exposed to an LLM can obtain approval from AICT, helping to ensure autonomous decisions never bypass enterprise governance. It works with AI Agent Studio, Claude, and so on. For more information, see [ServiceNow MCP Registry](https://www.servicenow.com/docs/RqTnkFqsAkvzQf3pPwwteg "Model Context Protocol (MCP) enables AI agents to connect to external systems and perform the required actions without any human intervention. ServiceNow MCP Registry is a private registry where large language models (LLMs), AI agents, and teams discover and connect to approved MCP servers, finding the best tool for every business problem, governed at every step. With these configurations, you can use any generative AI assistant to perform the required actions in ServiceNow."). ## MCP connector registration methods for OAuth 2.1 {#model-context-protocol-connector__section_jwq_lx2_mjc} Connect Hub supports two OAuth client registration methods for connecting to external systems: manual client registration and dynamic client registration. When the MCP registry does not include a connector for the system you need, you can define a custom connector. Connectors that are tested and approved by ServiceNow display a Certified by ServiceNow label in the registry. Dynamic client registration automates the OAuth application setup between Connect Hub and the third-party provider. Instead of requiring you to create an OAuth application manually, the provider and Connect Hub negotiate the client credentials automatically as part of the connection flow. When a connector may probably require manual setup, the registry entry displays the label Manual setup may be required next to the connector name. Dynamic client registration offers the following advantages over manual registration: * Credentials are managed by the provider and rotated automatically, reducing the risk of credential exposure. * The connection workflow is completed in fewer steps. {#model-context-protocol-connector__ul_y5h_q1f_mjc} Manual client registration requires you to create and configure an OAuth application in the third-party system's developer settings before connecting the MCP connector. During this process, the third-party system issues Client ID and a Client secret, which you provide to Connect Hub to authorize the connection. The MCP registry in Connect Hub includes connectors for a curated set of third-party systems. If the registry does not contain a connector for the system you need, you can define a custom connector by providing your own MCP endpoint URL and authentication configuration. ## Authentication methods for custom connectors {#model-context-protocol-connector__section_auth_methods} When creating a custom MCP connector, you select an authentication method that matches the security requirements of your external MCP server. The following authentication methods are supported: * **OAuth 2.1**: The recommended method for MCP connectors. Supports both Dynamic Client Registration, where the server auto-populates client details, and Manual Client Registration, where you supply the client credentials directly. * **API key**: Uses a static key issued by the third-party application to authenticate requests to the MCP server. * **Basic Authentication**: Authenticates using a username and password associated with the MCP server. {#model-context-protocol-connector__ul_dpy_pwk_mjc}