Studio 3T 2026.9 or later includes a built-in local MCP (Model Context Protocol) server that lets you connect your AI client of choice directly to your databases. Once connected, your AI can browse, query, and interact with your data without leaving your workflow.
You can use the Studio 3T MCP server with any MCP-compatible AI client, such as VS Code (GitHub Copilot), Cursor, Claude Code, Claude Desktop, Cline, or Gemini CLI.
Step 1: Enable the Studio 3T MCP server
- Open Studio 3T
- Go to Settings / Preferences → External Tools.
- Toggle on the Local MCP Server.
- The server now runs locally at
http://127.0.0.1:27117/mcp
Please note: if port 27117 is already in use on your machine, you can change it in Settings / Preferences → External Tools. Make sure to use the updated URL in your client configuration.
Step 2: Configure your AI client
With the server running, add it to your AI client’s configuration. Follow the instructions for your specific client below.
- Use ⌘ Shift P (macOS) or Ctrl Shift P (Windows/Linux) to open the Command Palette
- Search for MCP: Add Server and select it
- Select HTTP
- Paste
http://127.0.0.1:27117/mcpand hit Enter - Type studio3t when asked for a Server ID, then hit Enter.
- Choose whether to add this server globally or only for the current workspace.
After having set these options, your mcp.json will look like this:
{
"servers": {
"studio3t": {
"type": "http",
"url": "http://127.0.0.1:27117/mcp"
}
}
}
At this point, open the chat panel with ⌥⌘B or ⌃⌘ or Ctrl+Alt+B (Windows/Linux), or open the Command Palette and select Chat: Focus on Chat View.
Switch to Agent mode, then type #list_connections to confirm the Studio 3T tools are available.
For more information, see the VS Code official MCP documentation.
- Open Cursor → Settings → Cursor Settings.
- Go to the MCP tab.
- Click + Add new global MCP server.
Enter the following configuration and save:
{
"mcpServers": {
"studio3t": {
"url": "http://127.0.0.1:27117/mcp"
}
}
}
For more information, see the Cursor official MCP documentation.
Open your terminal and run the command below to add the MCP server for all Claude Code sessions (not just the current directory):
claude mcp add --transport http studio3t http://127.0.0.1:27117/mcp --scope user
This adds the server to ~/.claude/settings.json. You can also edit that file directly:
{
"mcpServers": {
"studio3t": {
"type": "http",
"url": "http://127.0.0.1:27117/mcp"
}
}
}
Useful commands for managing your MCP configuration:
| Command | Description |
| claude mcp list | List all configured servers |
| claude mcp get studio3t | Get details for the Studio 3T server |
| claude mcp remove studio3t | Remove the server |
For more information, see the Anthropic official MCP documentation.
- In VS Code, open the Cline sidebar and click the MCP Servers icon.
- Click Edit MCP Settings.
- Add the following entry to the config file:
{
"mcpServers": {
"studio3t": {
"type": "streamableHttp",
"url": "http://127.0.0.1:27117/mcp"
}
}
}
Save the file. Cline will automatically connect to the server.
For more information, see the Cline official MCP documentation.
Edit ~/.gemini/settings.json and add:
{
"mcpServers": {
"studio3t": {
"httpUrl": "http://127.0.0.1:27117/mcp"
}
}
}
For more information, see the Gemini CLI documentation.
For any other MCP-compatible client, add the Studio 3T server using this configuration (check your client’s docs for the exact config file location):
{
"mcpServers": {
"studio3t": {
"httpUrl": "http://127.0.0.1:27117/mcp"
}
}
}
Once configured, restart or refresh your client. You should see a successful connection and the Studio 3T tools listed. If the connection fails or no tools appear, double-check that the MCP server is enabled in Studio 3T (Settings / Preferences → External Tools), and that the port in your config matches the one set there.
Step 3: Start working with your data
Once connected, your AI client has access to the following Studio 3T tools:
| Tool | Description |
| list_connections | List your configured MongoDB connections |
| connect | Open a connection so other tools can operate on it |
| list_databases | List databases on the active connection |
| list_collections | List collections in a database |
| query | Run a find query and return results |
| explain_query | Explain how a query will be executed |
| analyze_schema | Describe the shape of a collection |
| list_indexes | List indexes on a collection |
| get_collection_statistics | Return size, document count, and related stats |
| assess_collection_health | Summarise indexing and shape consistency |
Here are some example prompts to get you started:
- “List all my database connections”
- “Connect to my local MongoDB and show me the collections in the orders database”
- “Find the last 10 documents in the orders collection where status is pending”
- “Analyze the schema of the customers collection and highlight any inconsistencies”
- “Explain how this query will be executed and suggest any index improvements”