kaggle-mcp-server 0.2.0

A full-featured MCP server wrapping the Kaggle API — 52 tools across competitions, datasets, kernels, models, benchmarks, and discussions

A full-featured MCP server wrapping the Kaggle API — 21 tools across competitions, datasets, kernels, models, and discussions.

English | 中文

Why kaggle-mcp?

Kaggle provides an official remote MCP server (https://www.kaggle.com/mcp). Here's how kaggle-mcp compares:

Feature	kaggle-mcp	Kaggle Official MCP
Total tools	21	~15
Discussions (search/list/detail/comments)	✅ 4 tools	❌
Dataset creation & file upload	✅	❌
Architecture	Local (stdio)	Remote HTTP
Install	uvx kaggle-mcp-server	npx mcp-remote required
No remote MCP dependency	✅	❌
Auth	API token	OAuth 2.0 / token

When to use kaggle-mcp: You need discussion tools, dataset creation/upload, or native stdio support without remote MCP dependency.

When to use the official MCP: You prefer OAuth 2.0 auth or want zero local installation.

Quick Navigation

Section	Description
Prerequisites	Kaggle API token setup
Installation	uvx / pip / source
Configuration	Claude Desktop, Claude Code, VS Code, Cursor
Tools (21)	Competitions, Datasets, Kernels, Models, Discussions
Debugging	MCP Inspector
Development	Local development setup

Prerequisites

A Kaggle API token is required. You can authenticate using either method:

Option A: API Token (recommended)

1. Go to https://www.kaggle.com/settings → API → Create New API Token
2. Set the environment variable:

export KAGGLE_API_TOKEN="KGAT_xxxxxxxxxxxx"

Option B: kaggle.json

Download the token file from Kaggle settings, it will be saved to ~/.kaggle/kaggle.json:

{"username": "your_username", "key": "your_api_key"}

Installation

Note: MCP servers are launched automatically by MCP clients (Claude Code, VS Code, etc.) — you don't need to run them manually in the terminal. The commands below are what the client uses under the hood.

Using uvx (recommended)

No installation needed. uvx will automatically download and run the server:

# Used by MCP clients internally; no need to run this yourself
uvx kaggle-mcp-server

Using pip

pip install kaggle-mcp-server

From source

git clone https://github.com/Galaxy-Dawn/kaggle-mcp.git
cd kaggle-mcp
uv sync

Configuration

Claude Desktop

Add to your claude_desktop_config.json:

Using uvx (recommended)

{
  "mcpServers": {
    "kaggle": {
      "command": "uvx",
      "args": ["kaggle-mcp-server"],
      "env": {
        "KAGGLE_API_TOKEN": "KGAT_xxxxxxxxxxxx"
      }
    }
  }
}

Using pip

{
  "mcpServers": {
    "kaggle": {
      "command": "python",
      "args": ["-m", "kaggle_mcp.server"],
      "env": {
        "KAGGLE_API_TOKEN": "KGAT_xxxxxxxxxxxx"
      }
    }
  }
}

Claude Code

claude mcp add kaggle -- uvx kaggle-mcp-server

Or add to your project's .mcp.json (not settings.json):

{
  "mcpServers": {
    "kaggle": {
      "command": "uvx",
      "args": ["kaggle-mcp-server"],
      "env": {
        "KAGGLE_API_TOKEN": "KGAT_xxxxxxxxxxxx"
      }
    }
  }
}

VS Code

Add to .vscode/mcp.json (note: the key is "servers", not "mcpServers"):

{
  "servers": {
    "kaggle": {
      "command": "uvx",
      "args": ["kaggle-mcp-server"],
      "env": {
        "KAGGLE_API_TOKEN": "KGAT_xxxxxxxxxxxx"
      }
    }
  }
}

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "kaggle": {
      "command": "uvx",
      "args": ["kaggle-mcp-server"],
      "env": {
        "KAGGLE_API_TOKEN": "KGAT_xxxxxxxxxxxx"
      }
    }
  }
}

Tip: If you already have KAGGLE_API_TOKEN in your shell environment (e.g. in .bashrc or .zshrc), you can omit the "env" block.

Tools (21)

Competitions (6)

Tool	Description
competitions_list	Search and list Kaggle competitions
competition_files	List data files for a competition
competition_download	Download competition data files
competition_submit	Submit predictions to a competition
competition_submissions	View submission history
competition_leaderboard	View leaderboard (top 20)

Parameter details

1. competitions_list — Search and list Kaggle competitions.

   • Inputs:
     • search (string, optional): Search term to filter competitions.
     • category (string, optional): Filter by category (e.g. featured, research, playground).
     • sort_by (string, optional): Sort order (latestDeadline, numberOfTeams, recentlyCreated).
     • page (integer, optional): Page number for pagination. Default 1.
   • Returns: List of competitions with title, ref, category, deadline, and team count.

2. competition_files — List data files for a competition.

   • Inputs:
     • competition (string, required): Competition URL suffix (e.g. titanic).
   • Returns: List of file names and sizes.

3. competition_download — Download competition data files. Returns download URL.

   • Inputs:
     • competition (string, required): Competition URL suffix.
     • file_name (string, optional): Specific file to download. Empty for all files.
   • Returns: Download URL.

4. competition_submit — Submit predictions to a competition.

   • Inputs:
     • competition (string, required): Competition URL suffix.
     • blob_file_tokens (string, required): Blob file token from upload.
     • message (string, required): Submission description message.
   • Returns: Submission result details.

5. competition_submissions — View submission history for a competition.

   • Inputs:
     • competition (string, required): Competition URL suffix.
   • Returns: List of submissions with date, score, status, and description.

6. competition_leaderboard — View competition leaderboard (top 20).

   • Inputs:
     • competition (string, required): Competition URL suffix.
   • Returns: Top 20 team names and scores.

Datasets (6)

Tool	Description
datasets_list	Search and list Kaggle datasets
dataset_files	List files in a dataset
dataset_download	Download dataset files
dataset_metadata	Get dataset metadata
dataset_create	Create a new dataset
file_upload	Upload a file to Kaggle

Parameter details

1. datasets_list — Search and list Kaggle datasets.

   • Inputs:
     • search (string, optional): Search term.
     • sort_by (string, optional): Sort order (hottest, votes, updated, active).
     • file_type (string, optional): Filter by file type (csv, json, sqlite, etc).
     • page (integer, optional): Page number. Default 1.
   • Returns: List of datasets with title, ref, size, and download count.

2. dataset_files — List files in a dataset.

   • Inputs:
     • owner (string, required): Dataset owner username.
     • dataset_slug (string, required): Dataset slug name.
   • Returns: List of file names and sizes.

3. dataset_download — Download dataset files. Returns download URL.

   • Inputs:
     • owner (string, required): Dataset owner username.
     • dataset_slug (string, required): Dataset slug name.
     • file_name (string, optional): Specific file. Empty for all.
   • Returns: Download URL.

4. dataset_metadata — Get dataset metadata.

   • Inputs:
     • owner (string, required): Dataset owner username.
     • dataset_slug (string, required): Dataset slug name.
   • Returns: Dataset metadata dictionary.

5. dataset_create — Create a new dataset. Use file_upload first to get file tokens.

   • Inputs:
     • owner (string, required): Owner username.
     • slug (string, required): Dataset slug.
     • title (string, required): Dataset title.
     • file_tokens (string, optional): Comma-separated file tokens from file_upload.
     • license_name (string, optional): License (e.g. CC0-1.0, CC-BY-SA-4.0). Default CC0-1.0.
     • is_private (boolean, optional): Whether dataset is private. Default true.
   • Returns: Creation result details.

6. file_upload — Upload a file to Kaggle and get a token for dataset_create.

   • Inputs:
     • file_name (string, required): File name (e.g. data.csv, config.json).
     • content (string, required): File content as text.
   • Returns: File token string.

Kernels (3)

Tool	Description
kernels_list	Search and list notebooks/kernels
kernel_pull	Get a notebook's source code
kernel_push	Push/save a notebook to Kaggle

Parameter details

1. kernels_list — Search and list Kaggle notebooks/kernels.

   • Inputs:
     • search (string, optional): Search term.
     • competition (string, optional): Filter by competition.
     • dataset (string, optional): Filter by dataset.
     • sort_by (string, optional): Sort order (hotness, commentCount, dateCreated, dateRun, relevance, voteCount).
     • page (integer, optional): Page number. Default 1.
   • Returns: List of kernels with title, ref, votes, and language.

2. kernel_pull — Get a notebook's source code.

   • Inputs:
     • user_name (string, required): Kernel owner username.
     • kernel_slug (string, required): Kernel slug name.
   • Returns: Kernel metadata and source code.

3. kernel_push — Push/save a notebook to Kaggle.

   • Inputs:
     • title (string, required): Notebook title.
     • text (string, required): Notebook source code.
     • language (string, optional): Language (python, r). Default python.
     • kernel_type (string, optional): Type (notebook, script). Default notebook.
     • is_private (boolean, optional): Whether notebook is private. Default true.
   • Returns: Push result details.

Models (2)

Tool	Description
models_list	Search and list Kaggle models
model_get	Get detailed model information

Parameter details

1. models_list — Search and list Kaggle models.

   • Inputs:
     • search (string, optional): Search term.
     • owner (string, optional): Filter by owner.
     • sort_by (string, optional): Sort order (hotness, downloadCount, createTime, updateTime).
     • page_size (integer, optional): Number of results per page. Default 20.
   • Returns: List of models with title and ref.

2. model_get — Get detailed information about a specific model.

   • Inputs:
     • owner (string, required): Model owner username.
     • model_slug (string, required): Model slug/name.
   • Returns: Model metadata dictionary.

Discussions (4)

Tool	Description
discussions_search	Search Kaggle discussions
discussions_list	List discussions for a competition/dataset
discussion_detail	Get discussion content by ID
discussion_comments	Get comments for a discussion

Parameter details

1. discussions_search — Search Kaggle discussions.

   • Inputs:
     • query (string, required): Search query string.
     • page_size (integer, optional): Number of results (max 50). Default 20.
   • Returns: List of discussions with ID, title, and votes.

2. discussions_list — List discussions for a competition or dataset.

   • Inputs:
     • competition (string, optional): Competition slug to filter.
     • dataset (string, optional): Dataset ref to filter.
     • page_size (integer, optional): Number of results (max 50). Default 20.
   • Returns: List of discussions with ID, title, and votes.

3. discussion_detail — Get discussion content by ID.

   • Inputs:
     • discussion_id (integer, required): Numeric discussion ID.
   • Returns: Discussion title, author, votes, forum, and body content.

4. discussion_comments — Get comments for a discussion.

   • Inputs:
     • discussion_id (integer, required): Numeric discussion ID.
   • Returns: Link to discussion comments page.

Debugging

You can use the MCP Inspector to debug the server:

npx @modelcontextprotocol/inspector uvx kaggle-mcp-server

The Inspector will provide a URL to access debugging tools in your browser.

Development

git clone https://github.com/Galaxy-Dawn/kaggle-mcp.git
cd kaggle-mcp
uv sync

Then configure the server in your MCP client using the local path, or test with MCP Inspector.

Contributing

Contributions are welcome! Please open an issue or submit a pull request on the GitHub repository.

License

This project is licensed under the MIT License. See the LICENSE file for details.