Report Abuse

Basic Information

MCPJungle is a self-hosted registry and gateway for Model Context Protocol (MCP) servers that centralizes discovery and management of tool servers used by AI agents. It provides a single source-of-truth where developers and organisations register, manage and discover MCP servers and the tools they expose. The project includes a server that exposes a unified /mcp gateway and a CLI client to register, list, invoke and deregister servers and tools. It supports streamable HTTP and STDIO transports, persistent storage via a default SQLite database or an external Postgres instance, Docker images for local and production deployment, and integration examples for MCP clients such as Claude and Cursor. The README highlights production features like authentication, access control and observability while noting the project is beta software.

Links

Categorization

App Details

Features
Central registry and unified gateway endpoint (/mcp) for all registered MCP servers. CLI client with commands to register, deregister, list tools, invoke tools, enable/disable tools and inspect usage. Supports registering servers via JSON config or CLI flags. Transports supported: streamable HTTP and STDIO; STDIO servers are run as subprocesses per tool call. Docker Compose quickstart and dedicated Docker images including a larger stdio image. Storage: default SQLite or configurable Postgres via DATABASE_URL. Authentication for streamable HTTP servers via bearer tokens. Production mode with admin initialization, ACLs and per-client access tokens. Canonical tool naming pattern <server>__<tool_name>. Notes on limitations: no long-lived connections to registered servers and OAuth not yet supported.
Use Cases
MCPJungle simplifies connecting AI agents and developer tools to diverse MCP servers by providing a single gateway and canonical tool namespace, reducing per-client configuration. Developers can register remote or local tool servers once and let multiple MCP clients discover and call those tools through the proxy. Organisations benefit from self-hosting, which keeps traffic on-premises, and from production features like access control, token-based client authentication and observability for auditing and governance. The provided CLI and Docker images make onboarding and automation easier, and integration examples for Claude and Cursor illustrate how existing MCP clients can be pointed to the gateway with minimal changes. The project is aimed at developers and teams building production-grade agent systems.

Please fill the required fields*