🚅 Windows/Linux Version Feature Roadmap!

[theJayTea]

Hey everyone! 👋

I hope you're doing well :D

I'm sorry I haven't been able to update the Windows/Linux version in a while - I had a ton of stuff going on (first my IB exams, and then some!).

I'm excited to share that I've now been able to resume development, and I'm sharing my To-Do list/feature roadmap!
Feel free to reply with any other feature suggestions you have. I'll also be going through all the GitHub Issues again when I can to add anything I missed to this list.

Expect an update in < a week with a significant amount of improvements (covering a ton of stuff from the list below!) :D

Upcoming Stuff (soon! in the next ~2 updates spanning this month):

• update to latest Gemini models

  • add custom Gemini model option!

• try to find a better hotkey detection API that doesn't force us to use only a few hotkeys

• per-button hotkeys

• per-button model choice

• per-button icons! (and enhance the default state + supply my set-up as an example/on GitHub)

• fix start on boot!

  • and show the start on boot checkbox in onboarding(!!!)

• fix ERROR_TEXT_INCOMPATIBLE_WITH_REQUEST not showing its popup

  • and "An error occurred while streaming" output when the Gemini safety filters kick in!?

• improve "Copy as Markdown"

• set mouse cursor to loading spinner when processing?

• encrypt the locally stored API key in config.json

• try to improve the current annoying restart-with-every-button-edit architecture!

• manually compile PyInstaller's bootloader to significantly reduce AV false-positives

• sign the exe (can get a cheap signed MSIX by releasing on MS Store?)

• diff tracking feature to easily see what the AI changed (maybe in the response window?)

• more feature suggestions from GitHub Issues

• various bug fixes and improvements
  [final: cleanup of solved GitHub issues; can archive some that are out of scope.]

---

Future Plan (by the next few updates, in a little over a month!)

• add run as admin check/confirmation to onboarding, settings, and also option editor GUI if Writing Tools is running from a protected system folder where elevation is required for it to write to its config file (eg: program files)

• implement localization (momokrono has built the entire framework)

• Support to remove reasoning token markers if needed (DeepSeek R1 & Gemini 2.0 Thinking already work!)
  (Issue #109)

• fix the model sometimes seeing 2 "user+model" pairs each time, and also that the first message in "Ask your AI..." isn't later shown to the model.

[menelic]

welcome back - looking forward to the new releases! Kindly include #111 on tracking changes/diff checking as you mentuoned in #184 . Maybe looking at Whisperng provides some opportunities to inegrate voice in a useful way.

[theJayTea]

Welp, looks like I'll need a few more days for the release. I've had some unforeseen personal stuff that came up.

@menelic, hey! Ah, added that cool request! :) It's something I need to brainstorm for a bit, so while it won't be coming in the imminent release, it's something that's certainly on the table.

[Ulf3000]

cool features but the current version crashes often. I hope thats resolved before everything else.