MCP Directory

Label Studio MCP Server

Manage Label Studio projects, tasks, and model predictions from your MCP client via the label-studio-sdk.

Unverified
stdio (local)
API key
Stale
Python
View repo 31 Website

Add to your client

Copy the config for your MCP client and paste it into its config file.

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "label-studio-mcp-server": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/HumanSignal/label-studio-mcp-server",
        "mcp-label-studio"
      ],
      "env": {
        "LABEL_STUDIO_API_KEY": "your_actual_api_key_here",
        "LABEL_STUDIO_URL": "http://localhost:8080"
      }
    }
  }
}

Requires `uv` (the Python package runner). Install it from https://docs.astral.sh/uv/ if `uvx` is not found.

Step-by-step guides: Add to Claude Desktop · Add to Cursor · Add to Windsurf

Before you start

  • A running Label Studio instance accessible from where the MCP server runs
  • An API key obtained from your Label Studio user account settings
  • uvx (uv) available to launch the server from the Git repository

About Label Studio MCP Server

An MCP server from HumanSignal (the makers of Label Studio) that exposes a Label Studio instance to MCP clients. It uses the official label-studio-sdk to manage labeling projects, import and inspect tasks and annotations, and attach model predictions, all driven from natural language or structured tool calls.

Tools & capabilities (10)

get_label_studio_projects_tool

Lists available projects (ID, title, task count).

get_label_studio_project_details_tool

Retrieves detailed information for a specific project (project_id).

get_label_studio_project_config_tool

Fetches the XML labeling configuration for a project (project_id).

create_label_studio_project_tool

Creates a new project with a title, XML label config, and optional settings. Returns project details including a URL.

update_label_studio_project_config_tool

Updates the XML labeling configuration for an existing project (project_id, new_label_config).

list_label_studio_project_tasks_tool

Lists task IDs within a project (up to 100).

get_label_studio_task_data_tool

Retrieves the data payload for a specific task (project_id, task_id).

get_label_studio_task_annotations_tool

Fetches existing annotations for a specific task (project_id, task_id).

import_label_studio_project_tasks_tool

Imports tasks from a JSON file (a list of task objects) into a project. Returns import summary and project URL.

create_label_studio_prediction_tool

Creates a prediction for a specific task. Requires the prediction result as a list of dictionaries in Label Studio format; optional model_version and score.

When to use it

  • Spin up a new Label Studio labeling project from a chat prompt with a given XML config.
  • Import a JSON file of tasks into an existing project and inspect the import summary.
  • Check labeling progress by listing tasks and retrieving annotations for a project.
  • Attach model-generated predictions to tasks to pre-label data for human review.

Security notes

Requires a Label Studio API key (obtained from your user account settings) and the URL of your Label Studio instance, both passed as environment variables. The server has full programmatic access to your projects, tasks, and predictions, so treat the API key as a secret and grant access only to a trusted instance.

Label Studio MCP Server FAQ

How do I provide my Label Studio credentials?

Set LABEL_STUDIO_API_KEY and LABEL_STUDIO_URL in the env block of the MCP server definition. The API key comes from your Label Studio user account settings, and the URL points to your running instance (e.g. http://localhost:8080).

Do I need a separately running Label Studio instance?

Yes. The server talks to an existing Label Studio instance via the label-studio-sdk, so you must have one running and reachable from where the MCP server runs.

Which MCP clients are supported?

Any client that supports an mcpServers config block. The README documents adding the entry to Claude Desktop (claude_desktop_config.json) or Cursor MCP settings, launched via uvx.

#label-studio#data-labeling#annotation#machine-learning#predictions#humansignal

Alternatives to Label Studio MCP Server

Compare all alternatives →
Memory (Knowledge Graph)
AI, Data & Knowledge
74k

Official MCP server providing persistent, file-backed knowledge-graph memory across sessions.

Verified
stdio (local)
No auth
TypeScript
9 tools
Updated 5 months agoRepo
Sequential Thinking
AI, Data & Knowledge
62k

Structured step-by-step reasoning tool for breaking problems into revisable thought sequences.

Verified
stdio (local)
No auth
TypeScript
1 tool
Updated 6 months agoRepo
AWS Knowledge MCP
AI, Data & Knowledge
7.0k

Fully managed remote server for AWS docs, blogs, What's-New and Well-Architected guidance — no key.

Verified
HTTP (remote)
No auth
Hosted
6 tools
Updated 5 months agoRepo

Compare Label Studio MCP Server with:

vs Memory (Knowledge Graph)vs Sequential Thinkingvs AWS Knowledge MCPvs MCP Server Chart