Kaggle MCP

Get the status of a dataset. Only for your own datasets. Parameters ---------- dataset : str Dataset URL suffix in format <owner>/<dataset-name> (use "kaggle datasets list" to show options) Returns ------- dict { "status": "success" | "error", "message": str, "dataset_status": dict (only if success) }

Return a list of datasets. Parameters ---------- sort_by: str How to sort the result. Valid options are: - 'hottest': Sort by trending/popularity - 'votes': Sort by number of votes - 'updated': Sort by last update date - 'active': Sort by active status - 'published': Sort by publication date file_type: str The format of datasets to filter by. Valid options are: - 'all': Show all file types - 'csv': Show only CSV files - 'sqlite': Show only SQLite databases - 'json': Show only JSON files - 'bigQuery': Show only BigQuery datasets license_name: s…

Initialize kernel metadata configuration in a specified folder. Parameters ---------- folder: str Path to the directory where kernel metadata will be created. Creates kernel-metadata.json with default configuration values. Returns ------- dict { "status": "success" | "error", "message": str, "data": dict (optional, typically empty) }

List files associated with a specific Kaggle kernel. Parameters ---------- kernel: str Kernel identifier in format 'owner/kernel-name'. Example: 'username/my-analysis'. page_token: str, optional Token for pagination through results. Default: None page_size: int, optional Number of files to return per page. Default: 20 Returns ------- dict { "status": "success" | "error", "message": str, "data": list of file metadata (if successful) }

Download the execution output files from a Kaggle kernel. Parameters ---------- kernel: str Kernel identifier in format 'owner/kernel-name'. path: str Local directory path where output files will be saved. force: bool, optional Whether to overwrite existing output files. Default: False. quiet: bool, optional Whether to suppress progress messages. Default: True. Returns ------- dict { "status": "success" | "error", "message": str, "kernel": str, "saved_path": str (only if success) }

Download a Kaggle kernel and its files to a local directory. Parameters ---------- kernel : str Kernel identifier in format 'owner/kernel-name'. path : str Local directory path where kernel files will be downloaded. metadata : bool, optional Whether to download kernel-metadata.json file. Default: False quiet : bool, optional Whether to suppress progress messages. Default: True Returns ------- dict { "status": "success" | "error", "message": str, "kernel": str, "saved_path": str (only if success) }

Push a Kaggle kernel and its metadata from a local folder. This tool validates the specified kernel directory, ensures the required metadata file exists, and then pushes the kernel to Kaggle using the Kaggle API. It returns structured status information instead of raising exceptions, allowing LLMs or external systems to interpret results programmatically. Parameters ---------- folder : str Path to the directory containing: - kernel-metadata.json - Kernel notebook or script files. timeout : int, optional Maximum execution time in seconds. Default is None (no timeout). Return…

Get the status of a Kaggle kernel. Parameters ---------- kernel: str Kernel identifier in format 'owner/kernel-name'. Returns ------- dict { "status": "success" | "error", "message": str, "kernel": str, "data": dict (kernel status info, if success) }

Search and list Kaggle kernels based on specified criteria. Parameters ---------- page: int, optional Page number of results to return. Default: 1 page_size: int, optional Number of results per page. Default: 20 search: str, optional Custom search query string. competition: str, optional Filter kernels to a specific competition. dataset: str, optional Filter kernels to a specific dataset in format 'owner/dataset-name'. user: str, optional Filter kernels to a specific username. mine: bool, optional If True, show only your own kernels. Default: False language:…

Create a new model on Kaggle using metadata from a local folder. Parameters ---------- folder : str Absolute path to the folder containing model-metadata.json and model assets. Returns ------- dict { "status": "success" | "error", "message": str, "data": dict (API response with model id, ref, url, etc.) }

Delete a model from Kaggle. Parameters ---------- model: str Model identifier in format 'owner/model-name'. confirmation: bool Must be True to confirm deletion. Returns ------- dict { "status": "success" | "error", "message": str, "model": str, "data": dict (API response if available) }

Retrieve metadata and details for a specific Kaggle model, and save it as a JSON file. Parameters ---------- model: str Model identifier in format 'owner/model-name'. Example: 'google/bert' or 'username/custom-model'. path: str, optional Directory path where the JSON file ('model-metadata.json') will be saved. Default: current working directory. Returns ------- dict { "status": "success" | "error", "message": str, "model": str, "data": dict (model metadata if success), "file_path": str (path to saved JSON file if success) } Si…

Initialize model metadata configuration in a specified folder. Parameters ---------- folder: str Path to the directory where model metadata will be created. Returns ------- dict { "status": "success" | "error", "message": str, "data": dict (optional, typically empty) } Side Effects ------------ Creates the model-metadata.json file with default configuration in the specified directory.

Create a new model instance on Kaggle using metadata from a local folder. Parameters ---------- folder : str Path to the folder containing model-instance-metadata.json and model assets. quiet: bool, optional Whether to suppress progress and status messages. Default: False dir_mode: str, optional How to handle directories in the model instance: - 'skip': Ignore directories (default) - 'zip': Compress directories and upload as zip files Default: 'skip' Returns ------- dict { "status": "success" | "error", "message": str, "data": dict (API r…

Delete a model instance from Kaggle. Parameters ---------- model_instance: str Model instance identifier in format 'owner/model-name/framework/instance-slug'. confirmation: bool Must be True to confirm deletion. Returns ------- dict { "status": "success" | "error", "message": str, "model_instance": str, "data": dict (API response if available) }

Retrieve details of a specific model instance from Kaggle and save as JSON. Parameters ---------- model_instance: str Model instance identifier in format 'owner/model-name/framework/instance-slug'. path: str, optional Directory path where 'model-instance-metadata.json' will be saved. Default: current working directory. Returns ------- dict { "status": "success" | "error", "message": str, "model_instance": str, "data": dict (metadata if success), "file_path": str (path to saved JSON if success) } Side Effects ------------ Creates '…

Initialize model instance metadata configuration in a specified folder. Parameters ---------- folder: str Path to the directory where model instance metadata will be created. Returns ------- dict { "status": "success" | "error", "message": str, "data": dict (optional, typically empty) } Side Effects ------------ Creates the model instance metadata file with default configuration in the specified directory.

Update an existing model instance using metadata from a local folder. Parameters ---------- folder: str Path to the folder containing model-instance-metadata.json and updated assets. Returns ------- dict { "status": "success" | "error", "message": str, "data": dict (API response with id, ref, url, etc.) }

Create a new version of an existing model instance. Parameters ---------- model_instance: str Model instance identifier in format 'owner/model-name/framework/instance-slug'. folder: str Path to folder containing model instance files and metadata. version_notes: str, optional Description of changes in this version. Default: ''. quiet: bool, optional Whether to suppress progress messages. Default: False. dir_mode: str, optional How to handle directories: - 'skip': Ignore directories (default) - 'zip': Compress and upload as zip files. Default: 'skip'. Returns --…

Delete a specific version of a model instance from Kaggle. Parameters ---------- model_instance_version: str Model instance version identifier in format 'owner/model-name/framework/instance-slug/version-number'. confirmation: bool Must be True to confirm deletion. Returns ------- dict { "status": "success" | "error", "message": str, "model_instance_version": str, "data": dict (API response if available) }

Download all files for a model instance version to a local directory. Parameters ---------- model_instance_version: str Model instance version identifier in format 'owner/model-name/framework/instance-slug/version-number'. path: str, optional Local directory path where version files will be downloaded. If not specified, downloads to current working directory. force: bool, optional Whether to overwrite files if they already exist locally. Default: False. quiet: bool, optional Whether to suppress progress and status messages. Default: True. untar: bool, optional Whet…

List all files contained in a specific model instance version. Parameters ---------- model_instance_version: str Model instance version identifier in format 'owner/model-name/framework/instance-slug/version-number'. page_token: str, optional Token for paginating through results when there are multiple pages. Default: None. page_size: int, optional Number of files to return per page. Default: 20. csv_display: bool, optional Whether to format output as comma-separated values instead of table format. Default: False. Returns ------- dict { "status": "success" | "er…

Update an existing model on Kaggle using metadata from a local folder. Parameters ---------- folder : str Absolute path to the folder containing model-metadata.json and model assets. Returns ------- dict { "status": "success" | "error", "message": str, "data": dict (API response with model id, ref, url, etc.) }

Search and list Kaggle models based on specified criteria. Parameters ---------- sort_by: str, optional Sort results criteria. Valid options: - 'hotness': trending/popularity - 'downloadCount': number of downloads - 'voteCount': number of votes - 'notebookCount': number of notebooks - 'createTime': creation date search: str, optional Search term to filter results. Default: None owner: str, optional Username or organization to filter models by. page_size: int, optional Number of results per page. Default: 20 page_token: str, optional Token for pagina…