Kei-naga/AIPaste

C#

0

5 commits

updated May 16, 2026

See the code

README

AIPaste

AIPaste is a Windows application for editing clipboard text with AI. Simply summon the app with a hotkey and provide instructions to easily modify your copied text. The result is then copied back to your clipboard for immediate use.

AIPaste demo

πŸ“ Overview

This tool uses a large language model (LLM) to revise and proofread text copied to your clipboard. It streamlines tasks like converting a casual message to a friend into a business email or correcting typos and grammatical errors in a document.

It runs in the background and can be called up at any time with a hotkey, so it doesn't interrupt your workflow. It supports both locally running LLMs and Google's Gemini model, allowing users to choose the AI model that best suits their needs.


✨ Key Features

  • Clipboard Integration: Automatically reads text from the clipboard to be edited by the AI.
  • Hotkey Support: Set a global hotkey to call the application at any time.
  • AI-Powered Text Editing: The AI revises, translates, summarizes, and performs other text edits based on user instructions.
  • Model Selection: Switch between a locally running LLM and the Google Gemini API.
  • Settings Menu: Easily configure hotkey combinations, API keys for AI models, and other settings through a GUI.
  • Auto-Startup: Configure the app to launch automatically when your PC starts.

πŸ› οΈ System Requirements

  • OS: Windows 11 or later
  • Framework: .NET 8.0
  • Platform: WinUI 3 / Windows App SDK

πŸš€ Installation

1. Manual Placement of LLamaSharp Backend

This application uses the LLamaSharp library, but some of its binary files must be placed manually. This step is necessary to work around a known issue with the library (issue #382).

Please download the binary files for the following libraries and place them within the Binary/LLamaSharpBackend directory according to the specified structure.

  • LLamaSharp.Backend.Cpu
  • LLamaSharp.Backend.Cuda11.Windows
  • LLamaSharp.Backend.Cuda12.Windows
  • LLamaSharp.Backend.Vulkan.Windows

Directory Structure:

Binary
 └── LLamaSharpBackend
   └── win-x64
     └── native
       β”œβ”€β”€ avx
       β”‚   β”œβ”€β”€ ggml.dll
       β”‚   └── ...
       β”œβ”€β”€ avx2
       β”œβ”€β”€ cuda11
       └── ...

2. Build

Open the AIPaste.sln file in Visual Studio 2022 or later and build the project.

3. Local packaged installation

In this repository, the Binary\LLamaSharpBackend files are already present, so you can usually skip the manual placement step above unless you removed them.

The packaged app is the supported startup path. Directly running bin\x64\Debug\...\AIPaste.exe is not expected to work reliably.

From the repository root, run the following in PowerShell or Developer PowerShell for Visual Studio:

Set-Location <repository-root>

msbuild `
  '.\AIPaste\AIPaste\AIPaste.csproj' `
  /t:Build `
  /p:Configuration=Debug `
  /p:Platform=x64 `
  /p:GenerateAppxPackageOnBuild=true `
  /p:AppxBundle=Never `
  /nologo

powershell -ExecutionPolicy Bypass -File `
  '.\AIPaste\AIPaste\bin\publish\AIPaste_1.0.0.0_x64_Debug_Test\Install.ps1' `
  -Force -SkipLoggingTelemetry

After installation, launch the app from the Start menu as AIPaste.


How to Use

  1. Copy the text you want to edit to your clipboard.
  2. Press the configured hotkey (default is Ctrl + Alt + C) to bring up the AIPaste window.
  3. In the input box at the bottom of the window, type your instructions for the AI (e.g., "Rewrite this as a business email") and press the execute button.
  4. The AI-generated text will appear in the "Generated Result" field.
  5. Press the "Copy" button to copy the generated text to your clipboard, ready to be pasted into other applications.

βš™οΈ Settings

On the settings page, you can customize the following options:

  • Hotkey: Change the key combination to summon the application.
  • Auto-Startup: Set whether AIPaste should start automatically when Windows boots up.
  • Model Selection:
    • Local LLM: Set the path to your local model file and configure settings like GPU usage.
    • Gemini: Set the API key for using the Google Gemini API.

πŸ“„ License

This repository is licensed under the MIT License. See LICENSE.txt for details.

Kei-naga/AIPaste

C#

0

5 commits

updated May 16, 2026

See the code

README

AIPaste

AIPaste is a Windows application for editing clipboard text with AI. Simply summon the app with a hotkey and provide instructions to easily modify your copied text. The result is then copied back to your clipboard for immediate use.

AIPaste demo

πŸ“ Overview

This tool uses a large language model (LLM) to revise and proofread text copied to your clipboard. It streamlines tasks like converting a casual message to a friend into a business email or correcting typos and grammatical errors in a document.

It runs in the background and can be called up at any time with a hotkey, so it doesn't interrupt your workflow. It supports both locally running LLMs and Google's Gemini model, allowing users to choose the AI model that best suits their needs.


✨ Key Features

  • Clipboard Integration: Automatically reads text from the clipboard to be edited by the AI.
  • Hotkey Support: Set a global hotkey to call the application at any time.
  • AI-Powered Text Editing: The AI revises, translates, summarizes, and performs other text edits based on user instructions.
  • Model Selection: Switch between a locally running LLM and the Google Gemini API.
  • Settings Menu: Easily configure hotkey combinations, API keys for AI models, and other settings through a GUI.
  • Auto-Startup: Configure the app to launch automatically when your PC starts.

πŸ› οΈ System Requirements

  • OS: Windows 11 or later
  • Framework: .NET 8.0
  • Platform: WinUI 3 / Windows App SDK

πŸš€ Installation

1. Manual Placement of LLamaSharp Backend

This application uses the LLamaSharp library, but some of its binary files must be placed manually. This step is necessary to work around a known issue with the library (issue #382).

Please download the binary files for the following libraries and place them within the Binary/LLamaSharpBackend directory according to the specified structure.

  • LLamaSharp.Backend.Cpu
  • LLamaSharp.Backend.Cuda11.Windows
  • LLamaSharp.Backend.Cuda12.Windows
  • LLamaSharp.Backend.Vulkan.Windows

Directory Structure:

Binary
 └── LLamaSharpBackend
   └── win-x64
     └── native
       β”œβ”€β”€ avx
       β”‚   β”œβ”€β”€ ggml.dll
       β”‚   └── ...
       β”œβ”€β”€ avx2
       β”œβ”€β”€ cuda11
       └── ...

2. Build

Open the AIPaste.sln file in Visual Studio 2022 or later and build the project.

3. Local packaged installation

In this repository, the Binary\LLamaSharpBackend files are already present, so you can usually skip the manual placement step above unless you removed them.

The packaged app is the supported startup path. Directly running bin\x64\Debug\...\AIPaste.exe is not expected to work reliably.

From the repository root, run the following in PowerShell or Developer PowerShell for Visual Studio:

Set-Location <repository-root>

msbuild `
  '.\AIPaste\AIPaste\AIPaste.csproj' `
  /t:Build `
  /p:Configuration=Debug `
  /p:Platform=x64 `
  /p:GenerateAppxPackageOnBuild=true `
  /p:AppxBundle=Never `
  /nologo

powershell -ExecutionPolicy Bypass -File `
  '.\AIPaste\AIPaste\bin\publish\AIPaste_1.0.0.0_x64_Debug_Test\Install.ps1' `
  -Force -SkipLoggingTelemetry

After installation, launch the app from the Start menu as AIPaste.


How to Use

  1. Copy the text you want to edit to your clipboard.
  2. Press the configured hotkey (default is Ctrl + Alt + C) to bring up the AIPaste window.
  3. In the input box at the bottom of the window, type your instructions for the AI (e.g., "Rewrite this as a business email") and press the execute button.
  4. The AI-generated text will appear in the "Generated Result" field.
  5. Press the "Copy" button to copy the generated text to your clipboard, ready to be pasted into other applications.

βš™οΈ Settings

On the settings page, you can customize the following options:

  • Hotkey: Change the key combination to summon the application.
  • Auto-Startup: Set whether AIPaste should start automatically when Windows boots up.
  • Model Selection:
    • Local LLM: Set the path to your local model file and configure settings like GPU usage.
    • Gemini: Set the API key for using the Google Gemini API.

πŸ“„ License

This repository is licensed under the MIT License. See LICENSE.txt for details.