# Playground

The **Playground** is an in-dashboard agent harness for testing MCP servers before distributing them. It is the last step of the [build lifecycle](/docs/ai-control-plane/mcp-gateway/building-servers#the-build-lifecycle) before a server reaches real clients, and it works for remote and tunneled servers too. Open it from **MCP Gateway > Playground** in the project sidebar.

## Access requirements

> Viewing this page requires the `mcp:connect`, `mcp:read`, or `mcp:write` scope. Running tools in the playground requires the `mcp:connect` scope, and editing saved playground or server configuration requires the `mcp:write` scope. The default [Admin role](/docs/ai-control-plane/org-admin/roles-and-permissions) includes full access; the default Member role can open and use the playground but cannot edit saved configuration.

## Testing a server

Pick any of the project's MCP servers from the **Select MCP** picker. Servers built from OpenAPI documents and functions and servers added from the catalog or by URL both appear in the same list; tunneled servers appear when tunneling is enabled for the organization. Authenticate against the server if it requires it, then chat with a model that calls the server's tools live.

While chatting:

- Mention tools directly by typing `@` followed by the tool name to steer which ones the model uses
- Click **Show Logs** to open the logs panel and see each tool call's request and response
- Adjust the configuration panel to change the server, the model, its temperature and maximum tokens, and, for built servers, the environment and authentication used for the session
- For built servers, open **Manage tools** to add tools from any of the project's sources or remove tools from the server without leaving the playground; clicking a tool opens its name, description, and annotations for editing
- Share a chat with the share button, which copies a link to the conversation

Servers added from the catalog or by URL show a read-only list of the tools the upstream server advertises. Tool curation, authentication, and environment settings for those servers live on the server's own page.

## When to use it

Use the playground to verify a new source produces working tools, debug authentication before distributing a server to real agents through its install page or a [plugin](/docs/ai-control-plane/mcp-gateway/plugins), and reproduce tool behavior reported by users without leaving the dashboard.