Clipboard Assistant

English | 繁體中文

Clipboard Assistant is a Windows desktop tool built with Qt 6 for managing clipboard content and extending custom functionality through a plugin mechanism.

Warnings

• This application was developed largely with assistance from Gemini and GPT.
• OpenAI / Gemini API keys are currently stored in plain text through QSettings; on Windows, this typically means they are stored in the Registry.

Project Goals

• Stay in the system tray for quick access to the main window and settings window
• Run a sequence of custom actions through hotkeys, for example:
  • Automatically copy selected content and process it with AI using a predefined prompt, such as translation
  • Capture content on the screen and process it with AI using a predefined prompt, such as translation
  • Process clipboard content and then either paste the result directly or just update the clipboard for later use
• Load additional feature modules through the Qt plugin mechanism

Usage Examples

Configure a plugin

To use OpenAI features, you need to create a corresponding account first:

1. Click Settings in the lower-left corner of the main window to open the settings dialog.
2. In Module Settings, select OpenAI Assistant on the left, then click Advanced Settings on the right.
3. In Manage OpenAI Accounts, click + in the lower-left corner to add an account, then enter the name, API key, model, and base URL before clicking OK.
   • It is recommended to click Test Connection first to verify the settings.

Create an action set

1. Click + in the lower-left corner to create a new action set.
2. Enter the action set name and configure a shortcut if needed.
   • You can enable Global to make it available from any application.
3. Adjust the startup and completion behaviors as needed.
4. Add the required functions from Pipeline Actions in the lower-left corner.
   • Multiple functions can be chained together.
   • Some functions provide additional options in Step Configuration.

Example configuration

With a configuration like this, pressing Ctrl + Shift + S will let the user capture a region of the screen, send the captured image to the configured OpenAI service for inference, and display the result.

The result looks like this:

If needed, multiple functions can also be chained in one action set for more advanced processing:

Included Features

Built-in assistant modules

The main application includes several built-in assistant modules that can be added directly to an action set, for example:

• RegEx Assistant: matches, extracts, or replaces text using regular expressions
• External App Assistant: launches external programs
  • Can pass clipboard text into command-line arguments
  • Can capture the program output as the result
  • Can be chained with existing CLI tools or used to launch applications with specific arguments
• Text Input Assistant: inserts fixed text or prompts for manual input while the pipeline is running

These built-in assistants can be chained together with plugin modules to form multi-step action set workflows.

Plugin-based assistant modules

Some more advanced features are currently implemented as plugins:

• OpenAIAssistant: a plugin integrated with the OpenAI API
  • Supports text, images, and streaming responses
  • In theory, supports all OpenAI API-compatible services
• GemmiAssistant: a plugin for Google AI services
  • Not thoroughly tested
• ScriptAssistant: a script-based plugin that uses JavaScript via QJSEngine
  • Each script must define a process(text) function
  • Standard JavaScript APIs such as text.toUpperCase() and JSON.parse() can be used

Main Application Features

• Separate main window and settings window
• Settings window options include:
  • Global hotkey for opening the main window
  • Launch at startup
  • List of loaded plugins
• Main window can:
  • Display the current clipboard content
  • Display user-defined action sets
  • Create, edit, delete, and reorder action sets
  • Import or export one or more action sets as JSON
  • Assign hotkeys to individual functions
• Each action set can chain multiple steps into a processing pipeline
• Action sets support advanced options such as auto copy, screen capture, completion behavior, auto close, and start hidden
• When a plugin runs, it receives the current clipboard content or captured content and returns the processed result; the main application decides how to output it, such as updating the clipboard, showing the output, or pasting directly

Build Environment

• Windows 11
• Visual Studio 2026
• Qt 6.11 (QtCore, QtGui, QtWidgets, QtNetwork, with some modules also using QML)

Open ClipboardAssistant.slnx to build the solution.

Repository Structure

• ClipboardAssistant/: main application
• OpenAIAssistant/: OpenAI plugin
• GemmiAssistant/: Gemini plugin
• ScriptAssistant/: JavaScript plugin
• Common/: shared headers
• doc/: screenshots and document resources used by the README
• licenses/: third-party license files
• ClipboardAssistant.slnx: Visual Studio solution file
• THIRD_PARTY_NOTICES.md: third-party component notices

Licensing and Third-Party Components

• Refer to the root LICENSE.txt for the project code license.
• When distributing binaries, include:
  • THIRD_PARTY_NOTICES.md
  • license files under the licenses/ directory
• Qt licensing and source information:
  • https://www.qt.io/licensing/
  • https://code.qt.io/
• It is recommended to deploy Qt via dynamic linking to avoid restricting users' rights to replace LGPL components.