This MCP server provides a Model Context Protocol (MCP) endpoint that lets you interact with a running Label Studio instance using the label-studio-sdk. It enables programmatic management of labeling projects, tasks, and predictions through natural language queries or structured MCP calls from clients. With this server you can create, update, list, and view project configurations; import tasks from files; list tasks within projects; retrieve task data and existing annotations; and attach model predictions to specific tasks. The server requires a running Label Studio instance and an API key, and is configured by supplying the Label Studio URL and API key (via environment variables or a client configuration file). It exposes tools for project management, task management, and predictions, such as get_label_studio_projects_tool, list_label_studio_project_tasks_tool, import_label_studio_project_tasks_tool, and create_label_studio_prediction_tool, all implemented using the official label-studio-sdk for reliable communication. Example use cases include creating a project, importing tasks from JSON, listing tasks, fetching task data/annotations, and adding predictions.