MCP / AI and media

Kaggle MCP: AI workflows

Kaggle is a platform for data science and machine learning competitions, offering datasets, notebooks, and a collaborative community. Connect it to Claude, Codex, Cursor or any MCP client through one FlyMy.AI MCP.

Your own API key20 actionsAI and media

$ YOUR NEXT RUN

Claude CodeCodexCursorAny agent

Use Kaggle MCP in your assistant

Paste this prompt into Claude Code, Codex, Cursor or any agent. It reads the guide and helps you connect.

Read https://flymy.ai/skill.md and https://flymy.ai/mcp/kaggle.md. Help me connect Kaggle MCP to this assistant. Then help me run this task: Use Kaggle to competition download files. Ask me for the required inputs first and show the result.
Open FlyMy.AIDownload brief

What Kaggle MCP can do

20 actions in the FlyMy.AI catalog, the most used first. Your agent finds them through search_tools and calls the one the task needs.

ActionWhat it does
# KAGGLE_COMPETITION_DOWNLOAD_FILESDownload competition data files.
# KAGGLE_CONFIG_INITInitialize Kaggle API client configuration.
# KAGGLE_DATASET_CREATECreate a new Kaggle dataset with full metadata.
# KAGGLE_DATASET_STATUSGet the status of a dataset upload or processing job.
# KAGGLE_DATASET_VERSIONCreate a new dataset version on Kaggle.
# KAGGLE_KAGGLE_COMPETITION_SUBMITSubmit an entry to a Kaggle competition.
# KAGGLE_KAGGLE_CONFIG_DIRRetrieve the directory of the Kaggle API configuration file.
# KAGGLE_KAGGLE_CONFIG_FILERetrieve the filename of the Kaggle API configuration file.
# KAGGLE_KAGGLE_CONFIG_KEYSList local Kaggle API configuration keys.
# KAGGLE_KAGGLE_CONFIG_PATHRetrieve local Kaggle API configuration file path.
# KAGGLE_KAGGLE_CONFIG_RESETReset local Kaggle CLI configuration to defaults.
# KAGGLE_KAGGLE_CONFIG_SETSet a Kaggle CLI configuration parameter.
Show all 20 actions
ActionWhat it does
# KAGGLE_KAGGLE_CONFIG_UNSETUnset a Kaggle CLI configuration parameter.
# KAGGLE_KAGGLE_CONFIG_VIEWView local Kaggle API configuration.
# KAGGLE_KAGGLE_DATASET_INITInitialize a dataset-metadata.json file in a local folder.
# KAGGLE_KAGGLE_DATASET_LIST_FILESList files in a Kaggle dataset.
# KAGGLE_KAGGLE_KERNEL_INITInitialize a kernel-metadata.json file in a local folder.
# KAGGLE_KAGGLE_KERNEL_OUTPUTDownload the output of a Kaggle kernel.
# KAGGLE_KERNELS_STATUSGet the status of a Kaggle kernel run.
# KAGGLE_LIST_DATASETSList Kaggle datasets with filters and pagination.

Ask your agent

Requests to paste as they are. Each one runs on the actions named under it.

“Use Kaggle to competition download files. Ask me for the required inputs first and show the result.”uses KAGGLE_COMPETITION_DOWNLOAD_FILES
“Use Kaggle to config init. Ask me for the required inputs first and show the result.”uses KAGGLE_CONFIG_INIT
“Use Kaggle to dataset create. Ask me for the required inputs first and show the result.”uses KAGGLE_DATASET_CREATE

Connect Kaggle in a minute

One MCP for every integration. Add it once, then every MCP in the catalog is a request away.

  1. 01

    Add FlyMy.AI

    Add the FlyMy.AI MCP to your assistant with one of the commands below and sign in. No key to copy.

  2. 02

    Add your Kaggle API Key

    Paste it into the Kaggle connection settings once; it stays with FlyMy.AI, never in a prompt or a chat.

    • Kaggle API Key - Your Kaggle API Key. Get this from your Account > API section (downloads as kaggle.json). Get API key →
  3. 03

    Ask

    Paste a request from this page. Every run shows what it cost.

Claude Code
claude mcp add --transport http flymyai https://mcp-agents.flymy.ai/mcp

Then run /mcp in Claude Code and sign in.

Codex
codex mcp add flymy --url https://mcp-agents.flymy.ai/mcp

Then run codex mcp login flymy to sign in.

claude.ai and Claude Desktop
https://mcp-agents.flymy.ai/mcp

Settings, Connectors, Add custom connector: paste this URL.

Pay per run

FlyMy.AI bills the run: the model and the tool's seconds. Calls go to your own Kaggle account, under its plan and limits. No seats and no minimum.

Go to the source

From the FlyMy.AI platform catalog of 2026-11-03: name, description, actions and setup as the platform lists them. Examples are requests to paste, not recorded runs.