A tiny, minimalist AI chat window. Requires an API (OpenAI-compatible).
- Small white text box, no border
- Always on top
- Click center to type
- Default Ctrl+Enter to send
- Right-click for settings
- Hidden from taskbar / Alt+Tab
- Drag edges to move/resize (long-press near the edge, then drag)
-
Install Python 3.7+
-
Download or clone this repo, then open a terminal in the repo directory
-
Install dependencies:
pip install -r requirements.txt
-
Start:
python ghostpad.py
-
Right-click → LLM Settings → set your API key, base URL, and model
-
Click center → type → Ctrl+Enter to send
-
LLM Settings — Configure your LLM provider: API key, base URL (OpenAI-style), model name, and basic params (e.g., temperature, max tokens). Works with OpenAI or any service that follows the same API format.
-
Set Hotkey — Set global shortcuts (e.g., show/hide window). Save after edits.
-
Start New Chat — Clear the current conversation and start a blank one. Note: history for this session will be cleared.
-
History — View the transcript for the current session.
-
Help — Opens this document.
-
Hide — Hides the text window. Use the configured show/hide hotkey to bring it back.
-
Exit — Exit. History will be cleared
No response / Error after entering
- Check your internet connection
- Check that your API key, base URL, and model are set correctly
Window won’t move / resize
- Make sure you are long-pressing near the window border (not the center)
- Hold for ~0.5s, then drag
Configuration edits are not saved
- Click Save after modifying settings (you may need to scroll down or manually enlarge the settings window to see the button)
- All settings from the LLM Settings tab are stored locally at:
C:\Users\<YourUsername>\.ghostpad\config.ini - No telemetry is collected; requests are sent only to your configured LLM provider.