Skip to content
Scalekit Docs
Talk to an Engineer Dashboard

Connect AI agents to DataRobot

Scalekit connector
Open markdown

The DataRobot connector lets your AI agent act in each user's DataRobot account. Each user connects their own DataRobot access token once, and Scalekit sends it with every call, so your agent never handles credentials. It comes with 53 tools.

Tools
53
What they doRead · write · destructive
31 · 15 · 731 read15 write7 destructive
Users sign in with

Setup

  1. Install the SDK

    Terminal window
    npm install @scalekit-sdk/node dotenv
  2. Set your credentials

    Add your Scalekit credentials to your .env file. Find values in app.scalekit.com > Developers > API Credentials.

    .env
    SCALEKIT_ENVIRONMENT_URL=<your-environment-url>
    SCALEKIT_CLIENT_ID=<your-client-id>
    SCALEKIT_CLIENT_SECRET=<your-client-secret>
  3. Create the DataRobot connection

    In AgentKit > Connections, create a DataRobot connection. The name you give it is the connection_name your code passes. See Configure connections.

  4. Authorize a user and make your first call

    quickstart.mts
    import { ScalekitClient } from '@scalekit-sdk/node'
    import 'dotenv/config'
    import { createInterface } from 'node:readline/promises'
    const scalekit = new ScalekitClient(
    process.env.SCALEKIT_ENVIRONMENT_URL,
    process.env.SCALEKIT_CLIENT_ID,
    process.env.SCALEKIT_CLIENT_SECRET,
    )
    const actions = scalekit.actions
    const connector = 'datarobot'
    const identifier = 'user_123'
    // Generate an authorization link for the user
    const { link } = await actions.getAuthorizationLink({ connectionName: connector, identifier })
    console.log('Authorize DataRobot:', link)
    const rl = createInterface({ input: process.stdin, output: process.stdout })
    await rl.question('Press Enter after authorizing...')
    rl.close()
    // Make your first call
    const result = await actions.executeTool({
    connector,
    identifier,
    toolName: 'datarobot_batch_predictions_list',
    toolInput: {},
    })
    console.log(result)
    Terminal window
    npx tsx quickstart.mts

    Each user opens the link once and enters their DataRobot credentials there. See Authorize a user for the full flow and statuses.

Tools

Pass the exact name to execute_tool
Try in PlaygroundRequest a tool
  • datarobot_batch_prediction_getGet the status, progress, row counts and settings of one batch prediction job.Read-only

    Get Batch Prediction Job

    Get the status, progress, row counts and settings of one batch prediction job. Returns an object with batchPredictionJobDefinition, created, createdBy, elapsedTimeSec, failedRows, hidden, id, intakeDatasetDisplayName, jobIntakeSize, jobOutputSize, and more. Use datarobot_batch_prediction_get to poll a job from datarobot_batch_prediction_create.

    Inputs

    prediction_job_idstringrequired
    Batch prediction job ID. Example: 5e3c94aff86f2d10692497c5
  • datarobot_batch_predictions_listList batch prediction jobs, filtered by status, deployment or date.Read-only

    List Batch Prediction Jobs

    List batch prediction jobs, filtered by status, deployment or date. Returns data[] (each with batchPredictionJobDefinition, created, createdBy, elapsedTimeSec, failedRows, hidden, id, intakeDatasetDisplayName) plus count, next, previous, totalCount. Use datarobot_batch_predictions_list to find jobs. Use datarobot_batch_prediction_get for one job.

    Inputs

    cutoff_hoursinteger
    Only jobs created at most this many hours ago. Example: 24
    deployment_idstring
    Only jobs for this deployment. Example: 5e3c94aff86f2d10692497b6
    end_date_timestring
    Latest creation time, ISO 8601. Example: 2026-11-15T00:00:00Z
    limitinteger
    Maximum number of results to return. Example: 20
    model_idstring
    Only jobs that used this leaderboard model. Example: 5e3c94aff86f2d10692497b8
    offsetinteger
    Number of results to skip. Example: 0
    order_bystring
    Sort order which will be applied to batch prediction list. Prefix the attribute name with a dash to sort in descending order, e.g. "-created". One of: created, -created, status, -status. Example: createdone of created-createdstatus-status
    start_date_timestring
    Earliest creation time, ISO 8601. Example: 2026-11-01T00:00:00Z
    statusstring
    Only jobs with this status, such as INITIALIZING, RUNNING, COMPLETED, ABORTED or FAILED. Example: COMPLETED
  • datarobot_dataset_features_listList the features (columns) of a dataset with their types and summary statistics.Read-only

    List Dataset Features

    List the features (columns) of a dataset with their types and summary statistics. Returns data[] (each with dataQualityIssues, dataQualityIssuesTypes, datasetId, datasetVersionId, dateFormat, featureType, id, isZeroInflated) plus count, next, previous, totalCount. Use datarobot_dataset_features_list for a dataset. Use datarobot_project_features_list for a project.

    Inputs

    dataset_idstringrequired
    Dataset ID, from datarobot_datasets_list. Example: 5e3c94aff86f2d10692497b7
    featurelist_idstring
    Only features in this featurelist. Example: 5e3c94aff86f2d10692497c1
    include_data_qualitystring
    If true, include detected data quality issues. Example: true One of: false, true.one of falsetrue
    include_plotstring
    If true, include histogram data. Example: false One of: false, true.one of falsetrue
    limitinteger
    Maximum number of results to return. Example: 20
    offsetinteger
    Number of results to skip. Example: 0
    order_bystring
    Sort field. One of: featureType, name, id, unique, missing, stddev. Example: nameone of featureTypenameiduniquemissingstddevmeanmedianminmaxdataQualityIssues-featureType-name-id-unique-missing-stddev-mean-median-min-max-dataQualityIssues
    search_forstring
    Only features whose name contains this text (case-insensitive). Example: age
  • datarobot_dataset_getGet one dataset with its size, row and column counts, processing state and source.Read-only

    Get Dataset

    Get one dataset with its size, row and column counts, processing state and source. Returns an object with categories, columnCount, createdBy, creationDate, dataEngineQueryId, dataPersisted, dataSourceId, dataSourceType, datasetId, datasetSize, and more. Use datarobot_dataset_get for one dataset. Use datarobot_dataset_features_list for its columns.

    Inputs

    dataset_idstringrequired
    Dataset ID, from datarobot_datasets_list. Example: 5e3c94aff86f2d10692497b7
  • datarobot_dataset_projects_listList the projects built from a dataset.Read-only

    List Dataset Projects

    List the projects built from a dataset. Returns data[] (each with id, url) plus count, next, previous, totalCount. Use datarobot_dataset_projects_list to see where a dataset is used.

    Inputs

    dataset_idstringrequired
    Dataset ID, from datarobot_datasets_list. Example: 5e3c94aff86f2d10692497b7
    limitinteger
    Maximum number of results to return. Example: 20
    offsetinteger
    Number of results to skip. Example: 0
  • datarobot_dataset_versions_listList the versions of a dataset.Read-only

    List Dataset Versions

    List the versions of a dataset. Returns data[] (each with categories, columnCount, createdBy, creationDate, dataPersisted, datasetId, datasetSize, isDataEngineEligible) plus count, next, previous, totalCount. Use datarobot_dataset_versions_list to find a datasetVersionId.

    Inputs

    dataset_idstringrequired
    Dataset ID, from datarobot_datasets_list. Example: 5e3c94aff86f2d10692497b7
    categorystring
    Only versions with this category. One of: TRAINING, PREDICTION, SAMPLE. Example: TRAININGone of TRAININGPREDICTIONSAMPLE
    filter_failedstring
    If true, leave out versions whose import failed. Example: true One of: false, true.one of falsetrue
    limitinteger
    Maximum number of results to return. Example: 20
    offsetinteger
    Number of results to skip. Example: 0
    order_bystring
    Sorting order which will be applied to catalog list. One of: created, -created. Example: createdone of created-created
  • datarobot_datasets_listList datasets in the Data Registry that you can access.Read-only

    List Datasets

    List datasets in the Data Registry that you can access. Returns data[] (each with categories, columnCount, createdBy, creationDate, dataPersisted, datasetId, datasetSize, isDataEngineEligible) plus count, next, previous, totalCount. Use datarobot_datasets_list to find a dataset ID. Use datarobot_dataset_get for one dataset.

    Inputs

    categorystring
    Only datasets with this category. One of: TRAINING, PREDICTION, SAMPLE. Example: TRAININGone of TRAININGPREDICTIONSAMPLE
    filter_failedstring
    If true, leave out datasets whose import failed. Example: true One of: false, true.one of falsetrue
    is_deletedstring
    If true, return only soft-deleted datasets. Example: false One of: false, true.one of falsetrue
    limitinteger
    Maximum number of results to return. Example: 20
    offsetinteger
    Number of results to skip. Example: 0
    order_bystring
    Sorting order which will be applied to catalog list. One of: created, -created. Example: createdone of created-created
    use_case_idsstring
    Only datasets in these Use Cases (IDs, comma-separated). Example: 65a1b2c3d4e5f6a7b8c9d0e1
  • datarobot_deployment_accuracy_getGet accuracy metrics for a deployment over a period, based on actuals uploaded to DataRobot.Read-only

    Get Deployment Accuracy

    Get accuracy metrics for a deployment over a period, based on actuals uploaded to DataRobot. Returns data[] (each with baselineSampleSize, baselineValue, metricName, modelId, percentChange, sampleSize, value) plus batchIds, metrics, modelId, period. Use datarobot_deployment_accuracy_get to monitor model quality in production. Accuracy needs actuals and an association ID set on the deployment; check datarobot_deployment_capabilities_get.

    Inputs

    deployment_idstringrequired
    Deployment ID, from datarobot_deployments_list. Example: 5e3c94aff86f2d10692497b6
    baseline_model_idstring
    Model whose metric value is used as the baseline. Example: 5e3c94aff86f2d10692497b8
    endstring
    End of the period, ISO 8601 (default the next hour). Example: 2026-11-15T00:00:00Z
    metricstring
    Name of the metric to retrieve. Must be provided when using multiple modelId query params. One of: AUC, Accuracy, Balanced Accuracy, F1, FPR, FVE Binomial, FVE Gamma, FVE Multinomial, FVE Poisson, FVE Tweedie, Gamma Deviance, Gini Norm, Kolmogorov-Smirnov, LogLoss, MAE, MAPE, MCC, NPV, PPV, Poisson Deviance, R Squared, RMSE, RMSLE, Rate@Top10%, Rate@Top5%, TNR, TPR, Tweedie Deviance, WGS84 MAE, WGS84 RMSE. Example: AUCone of AUCAccuracyBalanced AccuracyF1FPRFVE BinomialFVE GammaFVE MultinomialFVE PoissonFVE TweedieGamma DevianceGini NormKolmogorov-SmirnovLogLossMAEMAPEMCCNPVPPVPoisson DevianceR SquaredRMSERMSLERate@Top10%Rate@Top5%TNRTPRTweedie DevianceWGS84 MAEWGS84 RMSE
    model_idstring
    Model in the deployment (default the champion). Example: 5e3c94aff86f2d10692497b8
    startstring
    Start of the period, ISO 8601 (default 7 days ago). Example: 2026-11-01T00:00:00Z
    target_classstring
    Target class to filter results for. Example: yes