Kaggle for AI agents
Kaggle is a platform for data science and machine learning competitions, offering datasets, notebooks, and a community. Product and engineering teams use it as a building block inside their own AI features. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.
What agents do in Kaggle.
Download a competition dataset
Download the data files for a competition to start building a model.
Check a dataset's metadata
Pull a dataset's metadata and description before deciding if it fits a project's needs.
Submit a competition entry
Submit a prediction file to a competition only after the user confirms the entry is final.
35 Kaggle actions, graded by risk.
Every Kaggle action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
15Look things up. Allowed by default.
- kaggle.get_modelGet Model Details
- kaggle.list_modelsList Kaggle Models
- kaggle.list_kernelsList Kaggle Kernels
- kaggle.list_datasetsList Kaggle Datasets
- kaggle.list_competitionsList Kaggle Competitions
- kaggle.get_model_instanceGet Model Instance Details
- kaggle.get_dataset_metadataGet Dataset Metadata
- kaggle.list_competition_filesList competition data files
- kaggle.list_kernel_output_filesList Kernel Output Files
- kaggle.list_model_instance_version_filesList Model Instance Version Files
- kaggle.download_datasetDownload Kaggle Dataset
- kaggle.download_dataset_fileDownload Kaggle Dataset File
- kaggle.download_competition_fileDownload competition file
- kaggle.view_competition_leaderboardView competition leaderboard
- kaggle.download_competition_leaderboardDownload competition leaderboard
write
19Create and change things. Allow, or ask the user first.
- kaggle.config_dirGet Kaggle Config Directory
- kaggle.config_setSet Kaggle Configuration
- kaggle.config_initInitialize Kaggle Configuration
- kaggle.config_keysList Kaggle Configuration Keys
- kaggle.config_pathGet Kaggle Config Path
- kaggle.config_viewView Kaggle Configuration
- kaggle.kernel_initKaggle Kernel Init
- kaggle.pull_kernelPull Kernel Code
- kaggle.config_unsetUnset Kaggle Configuration
- kaggle.dataset_initKaggle Dataset Init
- kaggle.kernel_outputDownload kernel output
- kaggle.dataset_createDataset Create
- kaggle.dataset_statusGet Dataset Status
- kaggle.kernels_statusGet Kernel Status
- kaggle.dataset_versionCreate Dataset Version
- kaggle.competition_submitSubmit Competition Entry
destructive
1Delete, cancel or archive. Ask first, or deny outright.
- kaggle.config_resetReset Kaggle Configuration
Kaggle in three steps.
- 01Your users connect KaggleThey add their Kaggle api key on Arc0 Connect, under your brand. It goes straight into the vault.
- 02You set the rulesReads run, writes like “get Kaggle Config Directory” can wait for the user, and “reset Kaggle Configuration” can be denied outright.
- 03Any agent can actYour agent calls Kaggle through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('kaggle', { read: 'allow', write: 'ask', // config_dir destructive: 'deny', // config_reset }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How Kaggle connects.
Users add their Kaggle api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.
The same Kaggle connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →
- AUTH
- API key
- CREDENTIALS
- Per-tenant encrypted vault
- MODEL SEES
- Results only, never credentials
- AUDIT LOG
- Every call, on every plan
Use Kaggle from any agent.
ClaudeChatGPTCursor
Codex
VS CodeOpenAI Agents SDKClaude Agent SDK
Vercel AI SDK
Mastra
LangGraph
Kaggle and Arc0, answered.
Can I use Kaggle with Claude, ChatGPT or Cursor?
Yes. Connect Kaggle to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Kaggle actions you allow.
How do users connect Kaggle?
Users add their Kaggle api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.
Which Kaggle actions can my agent take?
35 in total: 15 read, 19 write and 1 destructive, such as “get Kaggle Config Directory”. Your policies decide which of them each agent may call.
Can I stop my agent from deleting things in Kaggle?
Yes. Actions like “reset Kaggle Configuration” are graded destructive. Set destructive actions to deny, or to ask so the user approves each one, and blocked calls still show up on the audit log.
Can my own backend call Kaggle too?
Yes. The same Kaggle connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug Kaggle into your agent.
Your users connect Kaggle once, under your brand. Your agent gets 35 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan