> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mcpjungle.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Mcpjungle Documentation

> Run all your MCP servers behind one endpoint

Mcpjungle is a self-hosted MCP gateway for developers and teams who want to manage multiple MCP servers without scattered client configurations, duplicated setup, or inconsistent access control.

Use it **locally** to keep your personal MCP setup clean, or run it as **shared infrastructure** for a team with centralized discovery, access control, and observability.

<p align="center">
  <img src="https://mintcdn.com/mcpjungle/fTOQgeQW87hF5GH_/images/mcpjungle-diagram-april-2026.png?fit=max&auto=format&n=fTOQgeQW87hF5GH_&q=85&s=308578b1ec3d132aa29b2a86ae6cf529" alt="mcpjungle product diagram" width="1536" height="1024" data-path="images/mcpjungle-diagram-april-2026.png" />
</p>

Instead of wiring every MCP server into every AI client, register your servers once in Mcpjungle and let Claude, Cursor, Codex, or your own agents connect to a single MCP endpoint.

## Use Cases

Mcpjungle is primarily built for:

* **Developers** who want one clean local MCP endpoint instead of per-client server configuration.
* **Teams** who want shared MCP infrastructure with centralized discovery, access control, and observability.

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Start the gateway, add your first MCP server and connect Claude - in under 5 minutes.
  </Card>

  <Card title="Core Concepts" icon="diagram-project" href="/core-concepts">
    Learn how the gateway, CLI, servers, MCP and enterprise features fit together.
  </Card>

  <Card title="Supported features" icon="table" href="/support-matrix">
    View transport support, MCP surface coverage, authentication, integrations, and operational support levels.
  </Card>

  <Card title="Installation" icon="download" href="/installation">
    Install the CLI locally and choose the right runtime path for your gateway deployment.
  </Card>
</CardGroup>

<Info>
  AI agents can also access the mcpjungle documentation using MCP - `https://docs.mcpjungle.com/mcp`.
</Info>

## Why Mcpjungle?

MCP is powerful, but managing many MCP servers gets messy fast.

Without a gateway:

* Every client needs its own MCP server configuration.
* Tools, prompts, and resources are scattered across different servers.
* Access control is duplicated or missing.
* Teams have no shared view of available MCP tools.
* Local setups become hard to reproduce.

Mcpjungle gives you a single control point:

<CardGroup cols={2}>
  <Card title="One endpoint for everything" icon="network-wired">
    Register your MCP servers once. Every AI client and agent connects to a single `/mcp` endpoint.
  </Card>

  <Card title="Self-hosted infrastructure" icon="server">
    Run Mcpjungle locally for yourself or on your own infrastructure for a team. Your MCP traffic, tool access, and operational data stay inside infrastructure you control.
  </Card>

  <Card title="Tool Groups" icon="layer-group" href="/guides/tool-groups">
    Expose curated subsets of tools so each client only sees the tools it actually needs.
  </Card>

  <Card title="Access Control" icon="shield-halved" href="/governance/access-control">
    In enterprise mode, issue per-client tokens and explicitly control which MCP servers each client can reach.
  </Card>
</CardGroup>

Start with a local setup. Scale to a shared team gateway when you need it.

<img src="https://mintcdn.com/mcpjungle/Fp2FwP_EhuKqtWTe/images/dashboard-ui.png?fit=max&auto=format&n=Fp2FwP_EhuKqtWTe&q=85&s=bc7e28ce2f8cf7c30a5d191aa9bab087" alt="Mcpjungle dashboard UI" width="1919" height="1035" data-path="images/dashboard-ui.png" />

## Quickstart

<Steps>
  <Step title="Start the server">
    Pull the Docker Compose file and start Mcpjungle with a persistent Postgres database:

    ```bash theme={null}
    curl -O https://raw.githubusercontent.com/mcpjungle/MCPJungle/refs/heads/main/docker-compose.yaml
    docker compose up -d
    ```
  </Step>
</Steps>

<Card title="Continue with the Quickstart guide" icon="arrow-right" href="/quickstart">
  Follow the full quickstart to install the CLI, register your first MCP server, and connect your AI client.
</Card>

## Explore the docs

<CardGroup cols={2}>
  <Card title="Servers and tools" icon="server" href="/guides/register-http-servers">
    Register remote and STDIO-based MCP servers, curate what they expose, and manage prompts and groups.
  </Card>

  <Card title="Deploy and operate" icon="cloud" href="/deployment/production">
    Run Mcpjungle in local or enterprise mode with the right database, runtime, and observability setup.
  </Card>

  <Card title="Access and governance" icon="lock" href="/governance/overview">
    Understand enterprise mode, upstream authentication, client tokens, and user access.
  </Card>

  <Card title="Reference" icon="terminal" href="/reference/cli-overview">
    Use the CLI, HTTP API, config files, and environment variables with exact command and schema details.
  </Card>

  <Card title="Roadmap" icon="road" href="/roadmap">
    See what's being prioritized.
  </Card>
</CardGroup>