[PLEASE READ BEFORE ANSWERING TO PREVENT IRRELEVANT SUGGESTIONS FOR ME.]
I'm looking to improve my workflow on Linux and am searching for a specific type of speech-to-text application to run locally on my laptop.

My requirements are:

• 100% Local & Offline: All audio processing must happen on my own machine.
• High Accuracy: Quality should be on par with a good Whisper model. I'm not interested in older models like VOSK, as their accuracy doesn't meet my needs.
• Key Use Cases: My main goals are to dictate notes directly into my "Second Brain" style notes app and to send longer prompts to Large Language Model interfaces without breaking my flow.
• System-Wide Integration: This is the most crucial part. I want to press a hotkey and dictate directly into any active application (my code editor, a browser, a document, etc.).

For context, I use Speechnotes all the time because it supports models like tiny-faster-whisper, which is very fast and works perfectly for my use case. The problem is purely its workflow—having to transcribe in one window and then constantly copy-paste the text is exactly the process I want to eliminate.

My goal is to find a seamless solution that works like native OS dictation but is powered by modern, local models. Many Whisper UIs I've found are excellent but seem to have the same limitation. The paid options are too expensive for what they are, which is why I'm focused on finding a great offline version.

Does a tool like this exist for Linux? What are you all using to achieve this kind of workflow?

Thanks for any help!