Disclaimer
This application is completely free to use and does not charge any fees. It utilizes OpenAI's Whisper service, meaning you pay OpenAI directly for usage. Whisper operates on a pay-as-you-go model, which is typically more cost-effective than other services that require prepaid quotas. You can monitor your usage and costs easily through your OpenAI dashboard.

Usage Instructions
To get started, add your OpenAI API key to the config.json file. Starting from version 1.0 you will have a pop-up window asking you for a key. You don't need to manually put it into the JSON file.
Please note this window will pop up after you record a message for transcription in case your API key didn't work. We will add a clear identification in the next drop.

If you're unsure how to obtain your OpenAI API key, simply search for it in ChatGPT—it will guide you through the process step by step. Or use these instructions: https://help.openai.com/en/articles/4936850-where-do-i-find-my-openai-api-key

Permissions & Setup
The application intercepts keyboard shortcuts, requires microphone access and clipboard access, you may need to grant the necessary permissions in your operating system. Additionally, some antivirus programs might block these functionalities, so you may need to whitelist the app in your antivirus settings to ensure it works properly.

Keeping the App Easily Accessible
For a smoother experience, make sure the app is always visible in the taskbar. To do this:
 - Right-click on the taskbar.
 - Select Taskbar settings.
 - Scroll to Other system tray icons and find the app.
 - Toggle the app to Always show in the tray bar.
Starting from version 1.0 there's an additional pop-up with a wave animation that makes it clear that you're recording.

Pricing Information

Privacy Disclaimer
Your data remains fully local—nothing is uploaded anywhere. Your transcription history is stored on your computer and is accessible via right-click menu. The recorded audio files are deleted after transcription is completed.

Disclaimer
This is not the final version of the product, but it is fully functional and ready to use. If you encounter any issues or have questions, feel free to reach out via the website or contact the author. www.bitsmarter.me

Feedback & Contact
We'd love to hear your feedback! Feel free to visit bitsmarter.me to share your thoughts or get in touch with the author.

release 1.4 
	- added local whisper API transcription option
	- recommended to use with https://github.com/ahmetoner/whisper-asr-webservice docker running locally. configure default local API URL in config window: http://localhost:9000/asr?output=text 
	- further instructions and tutorials on bitsmarter.me

release 1.5
	- replaced whisper-asr support with speaches https://github.com/speaches-ai/speaches
	- run locally using https://speaches.ai/installation/
	- provide  http://localhost:8000 as local whisper path in settings (you can adjust ports as needed when running the docker locally)
	- improves performance, lower memory usage compared to asr, better with non english transcription
	- 4 model sizes supported:
		"large" => "Systran/faster-whisper-large-v3" - fast with CUDA, best precision, especially noticeable with multi-lingual support - definitely recommended if running on CUDA
    	"medium" => "Systran/faster-whisper-medium", - faster, still very good for english and other languages
        "small" => "Systran/faster-whisper-small", - faster, less precise, good for english, worse for other languages (still tolerable)
        "tiny" => "Systran/faster-whisper-tiny", - fastest, less precise - decent with english, bad with other languages

release 1.6
	- improved microphone persistence logic to match by device name instead of index for better reliability
	- fixed exceptions that occurred on first recording when no previous audio file existed
	- added model size persistence in config.json file to remember user's preferred model selection
	- added model size estimations display in settings to help users choose appropriate model
	- fixed race condition exceptions when accessing wave audio files during recording

release 1.7
	- added External API support with HTTP Basic Authentication for secure remote transcription
	- new third tab "External API" in configuration window with username/password fields
	- SSL certificate validation bypass for external APIs (equivalent to curl -k functionality)
	- model size selection available for both local and external API configurations
	- improved configuration window layout: username/password fields on same line to save space
	- increased form height to accommodate all configuration options
	- fixed tab selection logic to prevent jumping between tabs during configuration
	- all external API settings (URL, credentials, model size, tab selection) are persistent
	- external API uses WAV files directly (like local API) for better compatibility
	- ready for nginx reverse proxy setups with HTTPS and basic authentication