Skip to main content
Glama
RapidMiner / Altair AI Studio MCP

RapidMiner / Altair AI Studio MCP

RapidMiner / Altair AI Studio MCP

An MCP server that lets Claude open, see, and control Altair AI Studio (the current name for RapidMiner Studio).

Claude ──MCP/stdio──▶ src/index.js (Node) ──HTTP 127.0.0.1:8765 + token──▶ Java agent inside AI Studio's JVM
                                                                           ├─ Swing UI control (screenshots, clicks, typing, menus, trees…)
                                                                           └─ RapidMiner API (process, operators, parameters, run, results, settings)

The agent runs inside AI Studio (loaded with -javaagent). That way:

  • screenshots and clicks work on GNOME Wayland, with no xdotool, ydotool or screen-capture permissions;

  • Claude reads the real component tree (buttons, trees, tables, dialogs) instead of guessing pixels;

  • processes can be built and configured exactly through RapidMiner's own API, with Groovy as an escape hatch.

Setup

npm install
./agent/build.sh                 # compiles agent/aistudio-mcp-agent.jar against your AI Studio jars
claude mcp add rapidminer -s user -- node "$PWD/src/index.js"

Using with OpenCode

Add the server to ~/.config/opencode/opencode.jsonc (or an opencode.json in your project), using the absolute path to src/index.js:

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "rapidminer": {
      "type": "local",
      "command": ["node", "/absolute/path/to/AIStudioMCP/src/index.js"],
      "enabled": true
      // "environment": { "RAPIDMINER_HOME": "/path/to/AI Studio" }
    }
  }
}

AI Studio is expected at ~/Downloads/Altair/RapidMiner/AI Studio 2026.1.1. If yours is elsewhere, set RAPIDMINER_HOME (for both agent/build.sh and the MCP server's environment).

Start AI Studio through the launcher (npm run launch, or let Claude call rm_launch). The launcher matches the official AI-Studio.sh and only adds the agent. Your installation is not modified. If you start AI Studio from its normal shortcut, the agent is not loaded and the tools can't connect.

Related MCP server: mcp-java-dev-tools

Tools

Group

Tools

Lifecycle

rm_launch, rm_status

Screen control

rm_windows, rm_screenshot, rm_ui_tree, rm_ui_find, rm_ui_get, rm_click, rm_type, rm_key, rm_select, rm_menu, rm_scroll, rm_drag, rm_window

Process

rm_process_structure, rm_process_xml, rm_process_set_xml, rm_process_new, rm_process_open, rm_process_save, rm_process_validate, rm_process_run, rm_process_stop, rm_process_results

Operators

rm_operator_search, rm_operator_info, rm_operator_add, rm_operator_get, rm_operator_set_parameters, rm_operator_remove, rm_operator_rename, rm_operator_enable, rm_operator_select, rm_connect, rm_disconnect, rm_arrange

Configuration

rm_settings_get, rm_settings_set, rm_log, rm_groovy

Tips

  • Find components by id (from rm_ui_tree/rm_ui_find) or by text / name / tooltip / class (class matches superclasses too, e.g. JTree).

  • rm_screenshot returns window coordinates scaled by scale. Pass the same scale to rm_click/rm_drag when clicking by x,y.

  • Some screens (e.g. the Welcome dialog) are custom-painted: use a screenshot plus an x,y click.

  • Port notation for rm_connect: inner ports of a subprocess are <Chain>:inner (the main process is Process:inner or process).

  • rm_window close needs an explicit window id. Closing the main window also needs confirm_exit: true.

Files

  • src/index.js: MCP server (Node, stdio)

  • agent/src/aistudiomcp/*.java: the Java agent (Agent HTTP server, Ui Swing control, Proc RapidMiner API, LogCapture)

  • scripts/launch-ai-studio.sh: starts AI Studio with the agent (it also fixes missing DISPLAY/XAUTHORITY and the UTF-8 locale when started from MCP)

  • examples/iris_decision_tree.rmp: a sample process built entirely through the MCP

Security

The agent listens on 127.0.0.1 only. Every request needs a random token that is regenerated each launch and written to ~/.AltairRapidMiner/aistudio-mcp.json (mode 600). rm_groovy can run arbitrary code inside AI Studio, so treat this MCP like shell access.

Available Tools

41 tools
rm_arrangeB

Auto-arrange operators in all subprocesses of the canvas.

Inputs: no named parameters.

rm_clickA

Click a component (id or text/name/tooltip/class query) or a point (x,y). Buttons are pressed via their action; other components get real mouse events. Supports right-click (context menus) and double-click.

Inputs: x, y, id, name, +14 more.

rm_connectB

Connect an output port to an input port. Use 'Process:inner' (or 'process') / ':inner' for a subprocess's inner ports, e.g. connect 'Decision Tree'.'model' -> 'Process:inner'.'result 1'.

Inputs: to_port (required), from_port (required), to_operator (required), from_operator (required), +2 more.

rm_disconnectC

Disconnect a port.

Inputs: port (required), operator (required), direction, subprocess.

rm_dragC

Drag with the mouse between two window points (e.g. move an operator or draw a connection in the process canvas).

Inputs: to_x (required), to_y (required), from_x (required), from_y (required), +5 more.

rm_groovyA

Escape hatch: run a Groovy script inside AI Studio with full API access. Bound vars: mainFrame (com.rapidminer.gui.MainFrame), process (com.rapidminer.Process), out (PrintWriter). Returns the script result.

Inputs: script (required), on_edt, timeout_seconds.

rm_keyB

Press a key or chord on the focused component (or target): e.g. ENTER, ESCAPE, TAB, DELETE, F11, ctrl+S, ctrl+shift+Z.

Inputs: keys (required), id, name, text, +7 more.

rm_launchB

Start Altair AI Studio (RapidMiner) with the control agent, or confirm it is already running. Waits until the main window is ready.

Inputs: wait_seconds.

rm_logC

Recent AI Studio / process log lines (filter with contains).

Inputs: lines, contains.

rm_menuB

Invoke a main-menu item by path, e.g. 'File > Save Process as' or 'View > Restore Default View'. Without path, lists the menus with shortcuts.

Inputs: path, depth.

rm_operator_addC

Add an operator by key (e.g. 'retrieve', 'concurrency:parallel_decision_tree', 'normalize') to the process or into a subprocess of a chain. Optionally name it and set parameters.

Inputs: key (required), name, parent, arrange, +2 more.

rm_operator_enableC

Enable or disable an operator.

Inputs: name (required), enabled (required).

rm_operator_getB

All parameters (value/default/range) and port connections of an operator in the process.

Inputs: name (required).

rm_operator_infoB

Documentation of an operator key: description, all parameters (type, default, range), ports, subprocesses.

Inputs: key (required).

rm_operator_removeC

Delete an operator from the process.

Inputs: name (required).

rm_operator_renameC

Rename an operator.

Inputs: name (required), new_name (required).

rm_operator_selectB

Select an operator in the GUI and show it (its parameters appear in the Parameters panel).

Inputs: name (required).

rm_operator_set_parametersC

Set operator parameters, e.g. {"criterion":"gain_ratio","maximal_depth":"10"}. List/enum parameters use RapidMiner's string encoding (see rm_process_xml for examples).

Inputs: name (required), parameters (required).

rm_process_newA

Create a new empty process (may prompt to save the current one - check rm_windows afterwards).

Inputs: no named parameters.

rm_process_openC

Open a .rmp process file from disk.

Inputs: path (required).

rm_process_resultsB

Results of the last run: each result port's type plus a data preview (rows/columns) or text.

Inputs: max_rows.

rm_process_runB

Run the current process in AI Studio and (by default) wait for it to finish, returning results previews and warnings.

Inputs: wait, max_rows, timeout_seconds.

rm_process_saveB

Save the current process (to its location, or to a new .rmp path).

Inputs: path.

rm_process_set_xmlA

Replace the current process with the given process XML (as in a .rmp file). Fastest way to build a whole process.

Inputs: xml (required), keep_location.

rm_process_stopB

Stop the running process.

Inputs: no named parameters.

rm_process_structureB

Readable outline of the current process: operators (name [key]), non-default parameters, connections, problems, nested subprocesses.

Inputs: no named parameters.

rm_process_validateA

List all setup problems (errors/warnings) in the current process.

Inputs: no named parameters.

rm_process_xmlA

Return the current process as RapidMiner process XML.

Inputs: no named parameters.

rm_screenshotA

Screenshot an AI Studio window (default: active) or a single component. Rendered from inside the app, so it works on Wayland. Coordinates in the image = window coordinates when scale=1.

Inputs: id, name, text, class, +6 more.

rm_scrollC

Mouse-wheel scroll over a component or point. amount>0 scrolls down, <0 up.

Inputs: x, y, id, name, +9 more.

rm_selectB

Select an item in a combo box, list, table, tree or tab pane by text (or item_index). For trees, value can be a path 'Samples > data > Iris'. double_click=true opens/activates it.

Inputs: id, name, text, class, +7 more.

rm_settings_getB

Read AI Studio preferences (Settings > Preferences). Give key for one value, or filter to list matching keys.

Inputs: key, filter.

rm_settings_setC

Change an AI Studio preference by key and save it.

Inputs: key (required), value (required), save.

rm_statusC

AI Studio version, current process location, run state, operator/problem counts, perspective.

Inputs: no named parameters.

rm_typeA

Type value into a text field (found by target query, or the focused one). clear=true replaces the content; enter=true presses Enter after.

Inputs: value (required), id, name, text, +7 more.

rm_ui_findB

Find components on all windows by text/name/tooltip/class. Returns ids and positions.

Inputs: name, text, class, exact, +2 more.

rm_ui_getB

Details of one component: full text, and for tables/lists/trees/combos/tabs their rows/items/tabs.

Inputs: id, name, text, class, +5 more.

rm_ui_treeB

Dump the Swing component tree of a window (or a subtree via id). Each line: [id] Class "text" name= tip= @x,y wxh (window coords). Use interactive_only to cut noise.

Inputs: id, window, max_depth, max_nodes, +1 more.

rm_windowB

Window actions: front, close (like clicking X; needs an explicit window id), maximize, resize.

Inputs: action (required), width, height, window, +1 more.

rm_windowsA

List AI Studio's open windows and dialogs (ids, titles, modal, active).

Inputs: no named parameters.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    C2.9/5.0

    Scored across 41 tools

    Disambiguation3/5

    Most tools are distinguishable by domain and action, but several clusters have close boundaries: rm_ui_tree/rm_ui_find/rm_ui_get, rm_click/rm_select/rm_menu, and rm_window/rm_windows. Descriptions help, but an agent must read them carefully to choose correctly, so overlap remains notable.

    Naming Consistency4/5

    All names use consistent snake_case with an rm_ prefix, and most follow domain_action patterns such as rm_process_run and rm_operator_add. There are minor deviations with verb-only or noun-only names like rm_click, rm_type, rm_window, and rm_windows, but the overall convention is readable.

    Tool Count2/5

    At 41 tools, the surface is far larger than the typical well-scoped range and includes many granular UI-automation and process-introspection operations. Several tools could be consolidated, making the set feel heavy for an agent to navigate.

    Completeness4/5

    The server covers process lifecycle, operator management, UI interaction, settings, logs, and even a Groovy escape hatch, so major workflows are supported. Minor gaps remain around repository/data browsing and explicit process close/delete operations, but these are mostly workable via existing tools.

    Maintenance

    ActivityMaintained
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      B
      quality
      D
      maintenance
      Сonnects to live JVMs via JDWP protocol, enabling LLM agents to autonomously debug Java applications — attach to a running process, pause threads, set breakpoints (including conditional), inspect stacks with auto-resolved object fields, evaluate methods, set variable values, and diagnose issues like deadlocks and hung queries.
      28
      12 PyPI
      1
      MIT
    • F
      license
      Not graded
      quality
      B
      maintenance
      Bridges agentic coding tools and live Java runtime behavior through a lightweight sidecar agent. Attaches directly to a running JVM to provide bytecode-level runtime signals for probe-verified inspection and deterministic debugging.
      2
      -