Firebot Commands V3

This is a program that can be used to trigger URLs on voice commands. It is designed to work with Firebot, but will work with anything that can be triggered through a URL. To use it in Firebot create a preset effects list, click "how to trigger from StreamDeck" and then use the URL provided there. You can also use an "On Firebot Start" event in Firebot to start any of the .exe's provided. If you want to edit the configuration use the GUI exe, but if you just want it to run in the background you can instead use the logic exe (for a completely free program) or the whisper exe (for improved processing).

This has much improved trigger word detection. The logic script will work as it always has, just with a better chance of correctly picking up on trigger words. I've added a lot of trigger words that will sound like "modbot" if you want to use that, but I've found that because it is not a real word the Google API often has difficulty with it, and "moderator" is much more accurate, and I have included that as a trigger word as well. You can edit the trigger words in the config file or the attached GUI. The trigger words start the script listening for a phrase, that you can edit in the URL mapping of the config file, and when it finds a phrase that matches what you said it will trigger the associated URL. I may edit this to work like the whisper exe that triggers just one URL but provides a transcript for firebot to process in the future, but for now I'm out of time I can spend on it.

The whisper script will instead use the OpenAi Whisper transcript model which you can get from here: https://platform.openai.com/settings/organization/api-keys. Whisper is not free, but it's very cheap at 0.006 cents per minute of transcription. The Whisper script uses the free Google transcript API that was used before to detect trigger words and then the Whisper API to more accurately transcribe the audio to a text file only when a trigger word has already been detected. This means it only triggers the paid transcription service when you actually want it to process a command. It then triggers exactly one URL that you can use to trigger Firebot to process the text file however you like. To use the whisper script or exe you will need to add an OpenAi API key and a suitable URL to trigger in the config file (this can be edited with the GUI)

The GUI is now different, instead of a full program that runs the trigger word detection and URL triggering it is now just a dedicated GUI for editing the config file and running the correct exe or .py script. It can handle both .exe and .py, so if you have python and the dependencies installed on your PC you can just use the GUI to run the python scripts, making it much easier to edit them yourself to do whatever you want.

If you wanted to run this without Firebot (it should be suitable for any program that you want to trigger with voice commands) there is now a setting in the config file that should turn that off.

While trigger detection is better it now has redundant processes to make sure a trigger word isn't detected, and also several systems to stop multiple URLs being triggered from one command. This means you need at least 5 seconds between trying commands, otherwise it will assume it's a duplicate command and refuse to trigger it.