mcp.run

Provides a centralized registry and HTTP gateway that aggregates multiple MCP servers (both public and private) into a single standardized endpoint. Acts as a protocol-compliant proxy that normalizes access to heterogeneous MCP server implementations, allowing clients to interact with multiple servers through one URL without managing individual server connections or authentication credentials.

Integrates with external identity providers via OIDC (OpenID Connect) protocol and supports OAuth 2.0 flows with automatic Dynamic Client Registration (DCR). Enables centralized user authentication and authorization without requiring manual OAuth app registration, allowing organizations to delegate identity management to existing IdP infrastructure (Okta, Azure AD, etc.).

Implements validation workflow that tests MCP server functionality and compatibility before approving submission to the registry. Performs automated checks on server schemas, tool definitions, and execution behavior to ensure quality and prevent broken or malicious servers from being exposed to users.

Provides reusable configuration profiles that standardize MCP server setup and deployment parameters. Enables administrators to define configuration templates that enforce organizational standards, reducing manual configuration overhead and ensuring consistent server deployment across the platform.

Implements role-based permission model that controls which users can submit MCP servers to the registry, approve server submissions, and access specific tools. Enforces governance gates through admin-controlled approval workflows, preventing unauthorized tool exposure and enabling fine-grained access policies based on user roles and organizational structure.

Enables HTTP webhook triggers that invoke automated tasks and tool executions within the mcp.run platform. Accepts incoming HTTP requests with task payloads, executes associated MCP tools, and returns results, providing event-driven automation without requiring direct API calls. Supports integration with external systems via standard HTTP webhooks for triggering complex workflows.

Provides persistent storage for saved prompts and tool combinations, allowing users to define reusable task templates that combine multiple MCP tools with predefined parameters. Enables execution of these templates on-demand, supporting workflow repeatability and reducing manual configuration overhead for common task patterns.

Captures and logs all tool executions, server access, and administrative actions in real-time, providing audit trails for compliance and operational visibility. Enables tracking of who accessed which tools, when, and with what parameters, supporting forensic analysis and compliance reporting requirements.

Provides emergency control mechanism to instantly disable all MCP tool access and deactivate running agents through a single administrative action. Enables rapid response to security incidents or policy violations by cutting off tool execution without requiring individual server shutdown or configuration changes.

Implements data loss prevention mechanisms that scan and filter tool outputs to prevent sensitive data exposure. Applies configurable DLP rules to tool execution results before returning to clients, blocking or redacting outputs containing patterns matching sensitive data categories (PII, credentials, etc.).

Supports both cloud-hosted (managed service) and self-hosted deployment models via containerized distribution (Docker implied). Enables organizations to deploy mcp.run within their own infrastructure for data residency, compliance, or operational control requirements, with unified feature parity between cloud and self-hosted variants.

Enables automated email delivery of task execution results and notifications through integrated email transport. Supports sending tool outputs, task completion status, and alerts via email without requiring external email service configuration, providing simple notification mechanism for non-interactive workflows.