Skip to main content

Sketch2App Workflow

Overview​

The flow of creating applications using Sketch2App can be summarized in the following sequential steps:

Step 1: Sketch input​

First, import a sketch to start creating your web application.

  • Import the sketch.
  • To recognize all the properties of the imported sketch, Sketch2App utilizes interconnected machine-learning models in its core pipeline.
  • During the inference process, an intermediate token called the DSL token is generated. This token contains crucial information about the sketch’s layout, coordinates, text, positions, shapes, and overall structure.
  • To convert the DSL tokens into the respective h2o-wave component codes, a compiler is employed. These component codes are then combined to create the web application.

Learn more about importing and processing Sketches in GenAI AppStudio

Step 2: Dataset input (optional)​

Secondly, you can opt to input your own dataset.

  • Uploading a dataset dynamically changes the web application data of the graphs and table based on the user-input CSV.

Learn more dataset input in GenAI AppStudio

Step 3: Document input (optional)​

The third step involves performing RAG on the uploaded documents.

  • Users have the option to upload a document, and GenAI AppStudio deploys the Retrieval Augmented Generation (RAG) technique on the provided document.

Learn more document input in GenAI AppStudio

Step 4: Trigger the Sketch2App pipeline​

The forth step involves triggering the Sketch2App pipeline. The pipeline is triggered to process the imported sketch, generate the corresponding DSL token, and finalize creating the application source code.

  • The generated source code serves as the foundation for running the web application. The generated application can be further refined and personalized using the built-in editing options within the application.
  • The capabilities include a Code Editor that allows you to make adjustments to the source code itself, a Wizard-based theme editor for adjusting the application’s visual elements, and an LLM-powered content generator to enhance the app’s content.
  • Sketch2App allows users to download the code and the application. This enables users to deploy and utilize the application in different environments according to their preferences.

Learn more about download and deployment in GenAI AppStudio

Step 5: Content generation and personalization​

Finally, explore how GenAI AppStudio integrates h2oGPT for content generation within Sketch2App.

  • After the component code is combined, the generated application is integrated with an LLM to facilitate content generation. While Sketch2App comes pre-configured with h2oGPT, users have the option to configure and utilize other LLM or GPT models according to their requirements.
  • The integration with LLM models offers the user the capability to generate images using text prompts.

Learn more about content generation and personalization in GenAI AppStudio

Feedback