DataRobot for AI agents
DataRobot automates machine learning model building, deployment, and monitoring, letting organizations turn large datasets into predictions without a full data science team. 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 DataRobot.
Check a project's training status
Check a project's status to see whether its model training run has finished.
Run a batch of predictions
Create a batch of predictions from a trained model against a new dataset.
Confirm before archiving a model package
Confirm before archiving a model package, since it stops being available for new deployments.
840 DataRobot actions, graded by risk.
Every DataRobot action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
462Look things up. Allowed by default.
- datarobot.get_quotasGet Quotas
- datarobot.get_datasetGet Dataset
- datarobot.get_projectGet Project
- datarobot.get_recipesGet Recipe
- datarobot.list_quotasList Quotas
- datarobot.list_statusList Status Jobs
- datarobot.get_notebookGet Notebook
- datarobot.list_versionList Version
- datarobot.get_batch_jobGet Batch Job
- datarobot.get_calendarsGet Calendar
- datarobot.get_genai_llmGet GenAI LLM
- datarobot.get_use_casesGet Use Case
- datarobot.list_datasetsList Datasets
- datarobot.list_projectsList Projects
- datarobot.get_custom_jobGet Custom Job
- datarobot.get_data_sliceGet Data Slice
write
299Create and change things. Allow, or ask the user first.
- datarobot.create_filesCreate Empty Files Catalog Item
- datarobot.update_groupsUpdate User Group
- datarobot.update_quotasUpdate Quotas
- datarobot.update_recipeUpdate Wrangling Recipe
- datarobot.create_commentCreate Comment
- datarobot.create_projectCreate DataRobot Project
- datarobot.update_commentUpdate Comment
- datarobot.update_projectUpdate Project
- datarobot.create_notebookCreate Notebook
- datarobot.create_use_caseCreate Use Case
- datarobot.update_calendarUpdate Calendar
- datarobot.update_datasetsUpdate Datasets (Bulk Action)
- datarobot.update_notebookUpdate Notebook
- datarobot.update_use_caseUpdate Use Case
- datarobot.create_custom_jobCreate Custom Job
- datarobot.create_deploymentCreate Deployment
destructive
79Delete, cancel or archive. Ask first, or deny outright.
- datarobot.delete_filesDelete Files
- datarobot.delete_quotaDelete Quota
- datarobot.delete_groupsDelete Multiple Groups
- datarobot.delete_recipeDelete Recipe
- datarobot.delete_statusDelete Status
- datarobot.delete_commentDelete Comment
- datarobot.delete_datasetDelete Dataset
- datarobot.delete_projectDelete Project
- datarobot.delete_use_caseDelete Use Case
- datarobot.delete_batch_jobDelete Batch Job
- datarobot.delete_calendarsDelete Calendar
- datarobot.delete_notebooksDelete Notebook
- datarobot.delete_custom_jobDelete Custom Job
- datarobot.delete_deploymentDelete Deployment
- datarobot.delete_user_groupDelete User Group
- datarobot.cancel_project_jobCancel Project Job
DataRobot in three steps.
- 01Your users connect DataRobotThey add their DataRobot api key on Arc0 Connect, under your brand. It goes straight into the vault.
- 02You set the rulesReads run, writes like “create Empty Files Catalog Item” can wait for the user, and “delete Files” can be denied outright.
- 03Any agent can actYour agent calls DataRobot through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('datarobot', { read: 'allow', write: 'ask', // create_files destructive: 'deny', // delete_files }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How DataRobot connects.
Users add their DataRobot 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 DataRobot 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 DataRobot from any agent.
ClaudeChatGPTCursor
Codex
VS CodeOpenAI Agents SDKClaude Agent SDK
Vercel AI SDK
Mastra
LangGraph
DataRobot and Arc0, answered.
Can I use DataRobot with Claude, ChatGPT or Cursor?
Yes. Connect DataRobot 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 DataRobot actions you allow.
How do users connect DataRobot?
Users add their DataRobot 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 DataRobot actions can my agent take?
840 in total: 462 read, 299 write and 79 destructive, such as “create Empty Files Catalog Item”. Your policies decide which of them each agent may call.
Can I stop my agent from deleting things in DataRobot?
Yes. Actions like “delete Files” 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 DataRobot too?
Yes. The same DataRobot connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug DataRobot into your agent.
Your users connect DataRobot once, under your brand. Your agent gets 840 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan