DataRobot

Create a new custom application source version in DataRobot. Use when you need to create a new version of an existing custom application source with optional file uploads, environment configuration, or based on a previous version.

Untrusted display-only summary

Tool to create a custom application source from a custom template. Use when you need to instantiate a new application source based on an existing custom template.

Untrusted display-only summary

Tool to create a new DataRobot custom job. Use when you need to define a custom execution task that runs arbitrary Python code in DataRobot's managed environment. Only the 'name' field is required; DataRobot applies sensible defaults for other fields.

Untrusted display-only summary

Tool to permanently delete a custom job. Use when you need to permanently remove a soft-deleted custom job and all its components. The custom job must be soft-deleted first before permanent deletion.

Untrusted display-only summary

Tool to create a hosted custom metric template for a DataRobot custom job. Use when defining how custom metrics should be collected and aggregated for observability. The template must specify whether metrics are model-specific, the aggregation type, and for numeric metrics, directionality and units.

Untrusted display-only summary

Tool to create a new DataRobot custom model for training or inference. Use when you need to register a custom model in DataRobot. For inference models, targetType and targetName are typically required (e.g., targetType='Regression' requires targetName).

Untrusted display-only summary

Tool to request logs from a deployed custom model. Use when troubleshooting failed prediction requests or debugging custom model behavior. Returns a status ID for polling - logs are generated asynchronously.

Untrusted display-only summary

Tool to clone an existing custom model in DataRobot. Use when you need to create a copy of a custom model for reuse or experimentation.

Untrusted display-only summary

Tool to create a new version for an existing DataRobot custom model. Use when you need to update a custom model with new code files, environment changes, or configuration updates. Creates either a major or minor version based on the isMajorUpdate parameter.

Untrusted display-only summary

Tool to start data disparity insight calculations for a DataRobot model. Use when you need to analyze data disparity between two classes for a specific feature. The calculation is asynchronous; poll the returned location URL to check job status and retrieve results.

Untrusted display-only summary

Create a Data Engine workspace state by executing a SQL query against DataRobot datasets. Use when you need to transform, join, or prepare data before modeling. The workspace state captures the query definition and can be used to generate new datasets or feed into projects.

Untrusted display-only summary

Create a dataset definition in DataRobot to define how to access and use a dataset from the AI Catalog. Use when you need to establish a reusable definition for a dataset that can be referenced in projects and deployments.

Untrusted display-only summary

Tool to create a chunk definition for a dataset definition in DataRobot. Use when you need to define how to partition a dataset into chunks for distributed processing or validation strategies.

Untrusted display-only summary

Tool to create a custom feature list within a dataset. Use when you need to define a specific subset of features for analysis or modeling.

Untrusted display-only summary

Tool to create a DataRobot dataset by uploading a file (CSV, Excel, etc.). Use when you need to upload data files to DataRobot's global catalog for modeling or prediction. Dataset creation is asynchronous - poll the statusId to monitor completion.

Untrusted display-only summary

Tool to create a feature transform on a DataRobot dataset. Use when you need to transform an existing feature into a new feature with a different type or extract date components. Feature creation is asynchronous - poll the returned location URL to check status.

Untrusted display-only summary

Tool to create a dataset from an external data source connector (database, S3, etc.). Use when you need to import data from a configured data source into DataRobot. The dataset creation is asynchronous - poll the returned location URL to check completion status.

Untrusted display-only summary

Tool to create a dataset from a DataRobot wrangling recipe. Use when you need to materialize a recipe into a reusable dataset. Creation is asynchronous - poll the location URL to check completion status.

Untrusted display-only summary

Tool to create a DataRobot dataset from a publicly accessible URL. Use when you need to import data from HTTP/HTTPS URLs into DataRobot's data catalog. Returns immediately with catalog and status IDs - dataset ingestion happens asynchronously.

Untrusted display-only summary

Tool to schedule a dataset refresh job in DataRobot. Use when you need to automate periodic data updates for a dataset. The schedule must be at least daily (hourly schedules are not supported).

Untrusted display-only summary

Tool to create dataset relationships in DataRobot by linking features between two datasets. Use when you need to enable advanced feature engineering across multiple datasets for modeling.

Untrusted display-only summary

Tool to create a new dataset version from the latest version of its data source. Use when you need to refresh a dataset with updated data from its original source. The dataset version creation is asynchronous - poll the returned location URL to check completion status.

Untrusted display-only summary

Tool to create a new version for an existing DataRobot dataset from a publicly accessible URL. Use when you need to update an existing dataset with new data from HTTP/HTTPS URLs. Returns immediately with status ID - version creation happens asynchronously.

Untrusted display-only summary

Tool to create a new dataset version from a specific previous version of its data source. Use when you need to refresh a dataset based on a particular historical version. The dataset version creation is asynchronous - poll the returned location URL to check completion status.

Untrusted display-only summary

Tool to create a new version of an existing DataRobot dataset by uploading a file (CSV, Excel, etc.). Use when you need to update an existing dataset with new data. Dataset version creation is asynchronous - poll the statusId to monitor completion.

Untrusted display-only summary

Create a new data slice in a DataRobot project to define a subset of data based on feature filters. Use when you need to analyze or model specific segments of your data (e.g., high-value customers, specific regions). Supports up to 3 filters with operators: eq, in, , between, notBetween.

Untrusted display-only summary

Tool to compute the number of rows available after applying a data slice to a dataset subset. Use when validating data slice filters or checking how many rows will be included in analysis for a specific data partition. Returns slice size and validation messages. Status 202 indicates successful validation.

Untrusted display-only summary

Tool to create a data stage in DataRobot. Use when you need to create a new data stage with a specified filename for data staging operations.

Untrusted display-only summary

Create a DataRobot deployment from a model package to enable real-time or batch predictions. Prerequisites: 1. First obtain a modelPackageId using LIST_MODEL_PACKAGES or CREATE_MODEL_PACKAGE 2. Optionally get predictionEnvironmentId using LIST_PREDICTION_ENVIRONMENTS 3. For managed SaaS, get defaultPredictionServerId…

Untrusted display-only summary

Tool to create a deployment custom metric from an existing custom job in DataRobot. Use when you need to track custom business or operational metrics for a deployment. The custom job must have an associated hosted custom metric template.

Untrusted display-only summary