# Register external models

> Register external models - Add an external model to the NextGen Registry.

This Markdown file sits beside the HTML page at the same path (with a `.md` suffix). It summarizes the topic and lists links for tools and LLM context.

Companion generated at `2026-05-06T18:17:10.046056+00:00` (UTC).

## Primary page

- [Register external models](https://docs.datarobot.com/en/docs/workbench/nxt-registry/nxt-model-directory/nxt-register-ext-models.html): Full documentation for this topic (HTML).

## Related documentation

- [NextGen UI documentation](https://docs.datarobot.com/en/docs/workbench/index.html): Linked from this page.
- [Registry](https://docs.datarobot.com/en/docs/workbench/nxt-registry/index.html): Linked from this page.
- [Models](https://docs.datarobot.com/en/docs/workbench/nxt-registry/nxt-model-directory/index.html): Linked from this page.
- [monitoring agent](https://docs.datarobot.com/en/docs/classic-ui/mlops/deployment/mlops-agent/index.html): Linked from this page.
- [Monitoring support for generative models](https://docs.datarobot.com/en/docs/classic-ui/mlops/monitor/generative-model-monitoring.html): Linked from this page.
- [Enable geospatial monitoring for a deployment](https://docs.datarobot.com/en/docs/workbench/nxt-console/nxt-monitoring/nxt-data-drift.html#enable-geospatial-monitoring-for-a-deployment): Linked from this page.
- [time series](https://docs.datarobot.com/en/docs/classic-ui/modeling/time/index.html): Linked from this page.
- [time step](https://docs.datarobot.com/en/docs/reference/glossary/index.html#time-step): Linked from this page.
- [multiseries models](https://docs.datarobot.com/en/docs/classic-ui/modeling/time/multiseries.html): Linked from this page.
- [deploy the model](https://docs.datarobot.com/en/docs/workbench/nxt-registry/nxt-model-directory/nxt-deploy-models.html): Linked from this page.

## Documentation content

To register an external model monitored by the [monitoring agent](https://docs.datarobot.com/en/docs/classic-ui/mlops/deployment/mlops-agent/index.html), add an external model as a registered model or version through Registry:

1. In theRegistry, on theModelstab, click+ Register a model(or thebutton when the registered model or version info panel is open): TheRegister a modelpanel opens to theExternal modeltab.
2. On theExternal modeltab, underConfigure the model, select one of the following options: Add a version to an existing registered modelCreate a new registered modelIncrement the version number and add a new version to the selected registered model.FieldDescriptionTargetThe dataset's column name that the model will predict on.Target typeThe type of prediction the model makes. Depending on the prediction type, you must configure additional settings:Regression: No additional settings.Binary: For a binary classification model, enter thePositive classandNegative classlabels and aPrediction threshold.Multiclass: For a multiclass classification model, enter or upload (.csv, .txt) theTarget classesfor your target, one class per line. To ensure that the classes are applied correctly to your model's predictions, the classes should be in the same order as your model's predicted class probabilitiesMultilabel: For a multilabel model, enter or upload (.csv, .txt) theTarget labelsfor your target, one label per line. To ensure that the labels are applied correctly to your model's predictions, the labels should be in the same order as your model's predicted label probabilitiesText generation:Premium feature. No additional settings. For more information, seeMonitoring support for generative models.Location:Premium feature. No additional settings. For more information, seeEnable geospatial monitoring for a deploymentAgentic Workflow:Premium feature. No additional settings. Agentic workflows can only be deployed on Serverless prediction environments. Agentic workflow deployments support monitoring service health, usage, custom metrics, and data exploration. They also support generating deployment reports.MCP Server:Premium feature. No additional settings. An MCP server is a tool that allows you to expose the DataRobot API as a tool that can be used by agents.Build environmentThe programming language used to build the model.Registered modelWhen saving as a version of an existing model, select the existing registered model that you want to add a new version to.Registered version nameAutomatically populated with the model name, date, and time. Change or modify the name as necessary.Registered model versionAssigned automatically. This displays the expected version number of the version (e.g.,V1,V2,V3) you create. This is alwaysV1when you selectRegister as a new model.Create a registered model and the first version (V1).FieldDescriptionRegistered model nameWhen registering a new model, enter a unique and descriptive name for the new registered model. If you choose a name that exists anywhere within your organization, a warning appears.Registered version nameAutomatically populated with the model name, date, and time. Change or modify the name as necessary.Registered model versionAssigned automatically. This displays the expected version number of the version (e.g., V1, V2, V3) you create. This is alwaysV1when you selectRegister as a new model.TargetThe dataset's column name that the model will predict on.Target typeThe type of prediction the model makes. Depending on the prediction type, you must configure additional settings:Regression: No additional settings.Binary: For a binary classification model, enter thePositive classandNegative classlabels and aPrediction threshold.Multiclass: For a multiclass classification model, enter or upload (.csv, .txt) theTarget classesfor your target, one class per line. To ensure that the classes are applied correctly to your model's predictions, the classes should be in the same order as your model's predicted class probabilitiesMultilabel: For a multilabel model, enter or upload (.csv, .txt) theTarget labelsfor your target, one label per line. To ensure that the labels are applied correctly to your model's predictions, the labels should be in the same order as your model's predicted label probabilitiesText generation:Premium feature. No additional settings. For more information, seeMonitoring support for generative models.Location:Premium feature. No additional settings. For more information, seeEnable geospatial monitoring for a deploymentAgentic Workflow:Premium feature. No additional settings. Agentic workflows can only be deployed on Serverless prediction environments. Agentic workflow deployments support monitoring service health, usage, custom metrics, and data exploration. They also support generating deployment reports.MCP Server:Premium feature. No additional settings. An MCP server is a tool that allows you to expose the DataRobot API as a tool that can be used by agents.Build environmentThe programming language used to build the model.
3. If registering atime seriesmodel, select theTime series modelcheckbox and configure the following fields: FieldDescriptionOrdering featureEnter the column in the training dataset that contains date/time values used by DataRobot to detect the range of dates (the valid forecast range) available for use as the forecast point.Date/time formatSelect the format of the model's forecast date and forecast point features, in GNU C library format. For example:%Y-%m-%dT%H:%M:%SZ (2012-07-31T04:00:00.000000Z)Forecast point featureEnter the column in the training dataset that contains the point from which you are making a prediction.Forecast unitSelect the time unit (seconds, days, months, etc.) of thetime step.Forecast distance featureEnter the column in the training dataset containing a unique time step—a relative position—within the forecast window. A time series model outputs one row for each forecast distance.Series identifier(Optional)Formultiseries models, enter the column in the training dataset that identifies which series each row belongs to.
4. If necessary, you can configure the followingOptional settings: FieldDescriptionRegistered version descriptionEnter a description of the business problem this model package solves, or, more generally, describe the model represented by this version.TagsClick+ Add tagand enter aKeyand aValuefor each key-value pair you want to tag the modelversionwith. Tags added when registering a new model are applied toV1.Training dataThe training data, uploaded locally or via theData Registry.Holdout dataThe holdout data, uploaded locally or via theData Registry. Use holdout data to set anaccuracy baselineand enable support for target drift and challenger models.Prediction columnIf you uploaded holdout data, enter the name of the column in the holdout dataset containing the prediction result.Model locationThe location of the model running outside of DataRobot. Describe the location as a file path, such asfolder1/opt/model.tar.
5. Once you've configured all required fields, clickRegister model. The model version opens on theRegistry > Modelstab. You candeploy the modelat any time.

**Q: Why provide holdout data?**

To set an accuracy baseline for external models (which enables target drift and challenger models when deployed), you must provide holdout data. This is because DataRobot cannot use the model to generate predictions that typically serve as a baseline, as the model is hosted in a remote prediction environment outside the application. Provide holdout data when registering an external model package and specify the column containing predictions.