Skip to content

Home Screen & Projects ​

This page documents the desktop launcher and local project behavior.

Home screen entries ​

The desktop app starts on a launcher-style home screen with the dAIve predict logo and six action cards.

Available entries:

  • New Project
  • Open Project
  • Only Prediction
  • Only Training
  • PDF Manual
  • Youtube Tutorial

These entries are the main desktop starting points. There is no separate cloud project dashboard like in the web app.

New Project ​

New Project opens a save dialog for a .dvp file path.

The dialog title reads "Choose location and name of the new project" and filters for dAIve predict Project (*.dvp).

After the path is chosen, dAIve predict:

  • creates the project directory
  • saves the initial project settings
  • opens Step 1: Training Data Import

This is the standard entry point when you want a full local project from scratch.

Open Project ​

Open Project opens a file dialog to select an existing .dvp file.

The dialog filters for dAIve predict Project (*.dvp) and All Files (*.*).

After opening, dAIve predict restores the saved state and navigates to the step that was active when the project was last saved.

You can also open .dvp files by double-clicking them in the file explorer (if file associations are set up).

Only Prediction ​

Creates a project shell that starts directly in Step 6: Prediction.

Steps 1–5 are disabled. Only the prediction step is active.

Use it when you already have a trained .dvm file and only want inference.

This path avoids reopening or rebuilding a full training project when only prediction is needed.

You can also launch prediction-only mode by opening a .dvm file directly from the command line or file explorer.

Only Training ​

Creates a project shell for the training workflow and opens Step 1: Training Data Import.

Use it when you want to move directly into data import and model creation without opening an older project first.

Application menu ​

The desktop app has a top-level application menu with several important actions.

Project menu ​

  • Save project — saves the current project to its .dvp file
  • Save project as... — opens a save dialog to save a copy under a new name
  • Close project — prompts to save unsaved changes before closing (Save and close / Discard and close / Cancel)
  • Exit — quits the application

Settings menu ​

  • Backend — opens a configuration window for the API URL, with a connection test
  • Proxy — opens a proxy configuration window

Data menu ​

  • Clear Invalid Data — deletes generated artifacts (models, predictions, analysis results) and resets steps 4–6 to their initial state

Help menu ​

  • PDF Manual — opens the documentation folder
  • Youtube Tutorial — opens the tutorial video in a browser
  • About dAIve predict — shows version information

Workflow navigation ​

Once inside a project, the app shows:

  • a stepper at the top with 6 steps (or fewer in training-only / prediction-only mode)
  • a bottom navigation bar with Previous, step tabs, and Next

Stepper steps ​

StepName
1Training Data Import
2Training Data Description
3Model Design
4Model Training
5Training Data Assessment
6Prediction

Steps show visual state indicators:

  • completed steps have a highlighted icon
  • the current step is visually prominent with a colored background
  • upcoming steps are grayed out and may be locked until prerequisites are met
  • steps with stale data show a yellow warning badge with a tooltip explaining the invalidation

Bottom navigation ​

  • Previous moves one step back (on Step 1 it becomes Save & Close Project)
  • Next moves one step forward (on the last step it becomes Save & Close Project)
  • step number tabs in the center allow direct navigation to any enabled step

For demo projects, the close action reads Close Demo Project instead.

Step prerequisites ​

Steps unlock based on progress:

  • Step 2 requires at least one input and one output selected in Step 1
  • Step 3 is always available after Step 2
  • Step 4 requires a model task to be configured in Step 3
  • Step 5 requires a trained model from Step 4
  • Step 6 requires a trained model (unless in prediction-only mode)

dAIve predict customer documentation for web app and desktop app