Getting Started with Screph

This section helps you install and launch Screph, choose an appropriate mode, and create the first task for AI on your own image, video or screen recording.

Installation

Download the installer for the Windows 10/11 x64 target from the Download page, review the system requirements, and follow the installation wizard.

Alternatively, you can use the source code from GitHub and install manually following the README instructions.

First Launch

Launch Screph from the shortcut after installation or from the Start menu. In the program you can open or capture a visual source, mark and connect elements, add text or voice descriptions, and save a structured task. Computer-vision methods are optional; the saved context can be continued in the agentic IDE.

For voice annotations (speech‑to‑text) you can choose an external provider or use offline mode with local Vosk models. Before the first working project, open the dependencies guide and prepare only the runtime components your task needs.

Choose a Program Mode

Open General Settings → Development Platform or the Program mode list in the Selector header. Choose one of the available modes: General, GUI Automation, Industrial Vision, or Autonomous Unmanned Vehicles. Switching updates the CV workspace and Selector capabilities together; save the project before moving between domain workflows.

By default, the wizard proposes %LOCALAPPDATA%\Screph, creates a Start menu shortcut and preserves user data during reinstall or uninstall unless you explicitly select removal. The Virtual HID driver and Arduino flashing are separate post-install actions; see Windows and permissions.

Start with General for free-form markup and GUI Automation for application screens. Industrial Vision and Autonomous Unmanned Vehicles already provide domain workspaces, preview, candidate review, and confirmed markup application, but they are not production-control or complete export systems. See Program Modes for details.

Interface Overview

The main Screph application window consists of several key areas:

  • Canvas: The central area where the current screenshot or image is displayed. This is where you will select elements and work with them.
  • Tools Panel: Contains buttons for selecting work modes (Select, Create, Move, Resize, Hierarchy) and other actions.
  • Info Panel: Displays properties of the selected element, element hierarchy and other useful information.
  • Main Menu: Located at the top of the window. Provides access to save/load projects, settings and help.
  • Voice Annotation Widget: Appears when the voice annotation feature is activated. Allows recording, stopping and viewing recognized text.

Main Steps

1. Source Data: Capture Tab or Open Image

Start with a visual source. The Capture tab can record a region, monitor or window, use a camera or URL stream, and combine multiple images, videos and directories into one timeline through Files / folders. A single still image can be loaded from the menu.

Quick link: the Capture tab guide.

2. Select a Canvas Tool

Use the Tools Panel to select a tool. These are local canvas operations, not the whole-program mode selected under Development Platform:

  • Select Mode: For selecting existing elements.
  • Create Mode: For drawing new elements on the canvas.
  • Move Mode: For moving elements.
  • Resize Mode: For resizing elements.
  • Hierarchy Mode: For creating parent-child relationships.

3. Working with Elements

Depending on the selected mode, click or drag on the canvas. Details in the "Element Selection" section of the User Manual.

4. Annotations

Add text and voice annotations to elements. If the task needs masks, contours, OCR, matching or other derived data, apply computer-vision methods; a proposed markup change is confirmed separately. Details in the "Annotations" section of the User Manual.

5. Save the Task and Continue

Use File → Save Project (Ctrl+S). File → Save Settings... (Ctrl+Shift+S) selects the environment, scope and post-save actions; Ctrl+O opens Project Manager. Saving creates a main project and a for_ai_agent context: continue the task in the Screph Code agentic IDE or a supported external tool, and use Automation Runtime separately for GUI automation. See Saving and Loading.

Settings and Hotkeys

Settings include selection options, interface, voice annotations and other parameters. Hotkeys are available for faster work - find the list in the settings section. Use the settings map; dedicated pages cover LLM/voice connections and Automation Runtime.

With this basic understanding you're ready to start. For detailed information, refer to the User Manual.