NextGen UI documentation > Applications > Application templates

Application templates¶

DataRobot offers various approaches for building applications; see the comparison table for more information.

Application templates provide a code-first, end-to-end pipeline for provisioning DataRobot resources. With customizable components, templates assist you by programmatically generating DataRobot resources that support predictive and generative use cases. The templates include necessary metadata, perform auto-installation of dependencies configuration settings, and seamlessly integrate with existing DataRobot infrastructure to help you quickly deploy and configure solutions.

Application templates contain three families of complementary logic. You can opt-in to fully-custom AI logic and a fully-custom front-end or utilize DataRobot's off-the-shelf offerings:

AI logic—Services AI requests, generates predictions, and manages predictive models.
App logic—Provides user consumption, whether via a hosted front-end or integrating into an external consumption layer.
Operational logic—Turns on all DataRobot assets.

Persistent storage in applications

DataRobot uses the key-value store API and file storage to provide persistent storage in applications. This can include user settings, preferences, and permissions to specific resources, as well as chat history, monitoring usage, and data caching for large data frames.

End-to-end walkthroughs

View end-to-end walkthroughs to learn how to make the most of application templates in DataRobot:

Walkthrough	Description
Talk to my Data Agent walkthrough	Learn how to build a Talk to my Data Agent application—from selecting the template in the Application Template Gallery to interacting with the AI agent.
Use a DataRobot deployment in an application template	Learn how to create a deployment from a RAG playground and vector database using financial jargon and then use that deployment to customize the Talk to my Data Agent application, allowing the agent to apply industry knowledge to its responses.

Available templates¶

The table below describes the available templates. Each template links to its respective GitHub repository. Note that, in addition to the templates that DataRobot provides, organization administrators can add custom application templates for their users to execute.

In addition to the templates listed below, you can review examples of DataRobot tasks you can perform with Pulumi.

Application template	Description
Forecast Assistant	Leverage predictive and generative AI to analyze a forecast and summarize important factors in predictions. The Forecast Assistant template provides explorable explanations over time and supports what-if scenario analysis. Example use case: store sales forecasting.
Guarded RAG Assistant	Build a RAG-powered chatbot using any knowledge base as its source. The Guarded RAG Assistant template logic contains prompt injection guardrails, sidecar models to evaluate responses, and a customizable interface that is easy to host and share. Example use cases: product documentation, HR policy documentation.
Predictive Content Generator	Generates prediction content using prediction explanations from a classification model. The Predictive Content Generator template returns natural language-based personalized outreach. Example use cases: next-best-offer, loan approvals, and fraud detection.
Predictive AI starter	Outlines a basic predictive AI deployment workflow in DataRobot using a Streamlit front-end. DataRobot recommends using this as a template for getting started with app templates, as it is easy to customize.
Talk to my Data Agent	Provides a talk-to-your-data experience. Upload a `.csv`, ask a question, and the agent recommends business analyses. It then produces charts and tables to answer your question (including the source code). This experience is paired with MLOps to host, monitor, and govern the components.
Talk to my Docs	Provides a talk-to-your-docs experience. Upload a document, ask a question, and the agent queries local and cloud-based document stores, providing accurate, context-aware information retrieval and analysis.
Cash Flow Forecasting	Outlines a basic development and prediction workflow for a late payment predictive model.
Revenue forecasting	Outlines a basic development and prediction workflow for a revenue forecasting predictive model.
Demand Planning	Outlines a demand planning predictive model development and forecasting workflow.
Delivery Planning	Outlines a delivery planning predictive model development and deployment workflow to predict the risk of delayed outbound customer deliveries.
Materials Planning	Outlines a materials planning predictive model development and deployment workflow to predict the risk of delayed inbound deliveries.

Open an app template¶

To access and configure an application template:

From the Workbench home page or Use Case directory, click Browse application templates to open the Application Gallery.

The Application Gallery displays the collection of available templates and provides a brief description of each. While the descriptions list example uses cases, each template can be applied to a variety of use case types.
Select a template to expand it and review its key features.

Note

For more information about a template, including the step-by-step procedures for running it, click Copy repository URL. Then, paste the URL into your browser to open the template in GitHub and view a detailed README that outlines the template's workflow.
Once you have selected an application template, click Open in a codespace. DataRobot then:
- Creates a Use Case for the template assets.
- Creates a codespace based off of the template, which includes all of the files necessary to execute the template's workflow.

Allow some time for the codespace to initialize. Then, you can work with the application template and follow its unique workflow to begin creating a variety of DataRobot resources.

In the codespace you have access to:

	Element	Description
1	Use Case	The Use Case that was created from the template. Click in the breadcrumbs to view the Use Case in the directory.
2	Codespace title	The name of the codespace created in a Use Case for the application template
3	The template README	The README file for the application template. It outlines steps for the local development workflow and critical steps for the codespace workflow that are slightly different (detailed in the "Setup for advanced users" section).
4	File browser	The codespace's file browser lists all of the files that are part of the application template. Use the browser to navigate to and open the README and notebooks executing the template's code.

How you work with an application template depends on where you work with it:

In a codespace, you open the terminal and execute the steps in the README that begin with creating an .env file.
Alternatively, clone the GitHub repository and work with the template in your local environment. Reference the template README documentation for the steps required to work with an application template locally.

Configure an app template with the CLI tool¶

After opening an application template in a DataRobot Codespace or GitHub, you can use the CLI tool to launch a more streamlined experience to setup and configure an application template. The CLI tool provides the following assitance:

Validates environment configurations and highlight missing or incorrect credentials.
Guides you through the setup process using clear prompts and step-by-step instructions.
Ensures the necessary dependencies and credentials are properly configured to avoid common configuration issues.

The CLI tool walks you through process of cloning the application template and successfully running it, resulting in a built application—all without having to consult the template's README or modifying the .env file directly.

To configure an application template with the CLI tool, open an application template in either a codespace or GitHub. This example uses a codespace, however, you can also run this on your local CLI.

Then, in the codespace, open Terminal from the left-hand navigation.

Local setup

If you are running the CLI tool locally, you must first install the binary for the CLI and the template's dependencies.

You can download and install the latest release by running the following command in your terminal:

# Linux or macOS
curl -fsSL https://raw.githubusercontent.com/datarobot-oss/cli/main/install.sh | sh

# Windows
irm https://raw.githubusercontent.com/datarobot-oss/cli/main/install.ps1 | iex

If you are using Homebrew, you may also:

brew install datarobot-oss/taps/dr-cli

Alternatively, you can install the CLI directly using its binary. When downloading and installing the binary directly, make sure to select the operating system and architecture that matches your local environment. For more information on dependencies, see the "Setup" section of the application template's README.

To launch a guided experience for setting up the application template, enter:

dr templates setup

The CLI automatically detects that you are inside of a template and generates a .env file.

To see a list of the available actions you can run, enter:

dr run

To run an action from the list of actions, enter:

dr run {action}

At the end of the guided setup, press E to edit the configuration, including API token, endpoint, and LLM information. To save and exit the configuration details, press Esc.

Add custom app templates to an organization¶

Organization admins can add custom application templates to their organizations, allowing users within the organization to build applications from them.

To add a custom application template:

Navigate to the Workbench home page and click Browse application templates to open the Application Gallery.
From the Application Gallery, click + Add template.

Configure the application template.

Field	Description
Template name	The name of the application template to be added to the Application Gallery.
Repository URL	A link to the repository that hosts the files that make up the application template.
Branch name	The Git branch stemming from the linked repository that contains the file contents for the application template.
Public toggle	Use this toggle to indicate whether the repository is public or not. Public repositories can be cloned without requiring authentication. If an application template is not marked as public, users are only able to copy the repository URL and will need to provide the appropriate credentials to clone the repository.
Readme	Upload a `README.md` Markdown file that outlines the purpose and workflow for the application template.
Image	Optional. Upload an image that represents the use case or workflow for the application template.
Tags	Provide tags to label the application template. To provide multiple tags, enter them in a comma-separated list.
Description	Optional. Provide a description that outlines what you would use the application template for in DataRobot.

After configuring the template, click Save template.

Once saved, the application template appears alongside other templates in the Application Gallery. Note that DataRobot-provided templates are indicated by the company logo (1), and admin-provided templates are indicated by the user icon (2).

Use a text generation NVIDIA NIM in an application template¶

Premium

The use of NVIDIA Inference Microservices (NIM) in DataRobot requires access to premium features for GenAI experimentation and GPU inference. Contact your DataRobot representative or administrator for information on enabling the required features.

GenAI application templates can integrate capabilities provided by NVIDIA AI Enterprise, a comprehensive, cloud-native software platform that accelerates data science pipelines and streamlines the development and deployment of production-grade AI applications. To use this integration, you can customize a DataRobot application template to programmatically generate a generative AI use case built on NVIDIA Inference Microservices (NIM).

To use an existing text generation model or deployment with these application templates, select one of the GenAI application templates from the Application Gallery. Then, you can make the following modifications, locally or in a DataRobot codespace, to customize the template to use a registered or deployed NVIDIA NIM.

In infra/settings_generative.py: Set LLM=LLMs.DEPLOYED_LLM.
In .env: Set exactly one of TEXTGEN_REGISTERED_MODEL_ID or TEXTGEN_DEPLOYMENT_ID.
In .env: Set CHAT_MODEL_NAME to the model name expected by the deployment. For NIM registered models and deployments, use "datarobot-deployed-llm".

With the template customized, you can proceed with the standard workflow outlined in the template's README.md.