So I made this thing where you can select the text you're typing (in any application) and press a shortcut key (Ctrl + . by default, not configurable yet) to include the selection into a prompt, send to an LLM (right now it's only OpenAI), and possibly insert the response back into the application you were in.
Also the same interface can be used to query indexed documents. Eventually I would have it watch a directory for changes so any documents added there are indexed automatically but right now indexing has to be done in settings screen.
I would appreciate your thoughts! You can test with your own OpenAI api key or you can register a test account and use my api key.
