
Beschreibung
GSXComms
What is GSXComms?
Talk to your GSX ground crewwith your voice instead of clicking through the menu. GSXComms is a Windows application that adds Ground Crew and life to GSX and MSFS airports.
Key Features:
- Supports both MSFS2024 and MSFS2020
- Speak with YOUR voice: You handle the Ground Comms, no copilot or AI doing things for you
- Fully FREE: 100% free, free of charge :)
- Fully offline: The app works perfectly fully offline (except the Mistral API option)
- The app has many accents, each one with a female and male version. The supported output accents are: American, Arabic, Australian, British, Chinese, Croatian, Filipino, Finnish, French, German, Greek, Indian, Irish, Italian, Japanese, Korean, Malay, Nigerian, Portuguese, Romanian, Russian, Scottish, Singaporean, Spanish, Swedish and Turkish
- GSXComms automatically locks your speech if your Ground Speed is above 15kts to prevent accidental commands during flight
- Multi language input: With the Mistral provider, you can speak in any language, and GSXComms will understand it. Note: The response will be in english
- Full GSX menu interaction: You can interact with all options in the GSX menu, for full interaction
- Presets: Set a codeword, and actions will be executed in sequence, one at a time (Set in settings)
- Match Speech to a GSX option: If your input is 75% (or more) similar to any GSX option, it will be automatically matched to a menu entry, to avoid typos or the voice recognition software just being weird (You can change this in settings)
- Top Most: Put GSXComms above any other window
- Show GSX Menu: See the GSX Menu without leaving GSXComms
- Static audio: For a realistic headset simulation, aiming at more immersion
- Auto-select operator: Automatically selects the best operator for your action
- Keyboard typing: Type the command you want to execute it, instead of speaking
& Much more!
All audio files were generated using ElevenLabs
Installation:
- Extract the downloaded zip to anywhere on your PC
- Run GSXComms.exe
- The launcher will automatically set up the necessary environment and download the dependencies.
- Once the window opens, bind your Push-to-Talk (PTT) key in the settings, and you are ready to fly!
Requirements:
- Microsoft Flight Simulator (2020 or 2024)
- GSX (FSDreamTeam).
- Python, 3.12.15 or newer, up to any 3.14.x (3.15 is not supported)
Update Instructions:
- Backup your settings (user_settings.json and Accents folder)
- Delete your old GSXComms folder
- Replace it with the new folder
- Re-add the user_settings.json and the Accents folder
Notes:
When using GSXComms, make sure that the GSX icon in the toolbar is NOT selected. If it is, the menu might keep popping up in the screen
Python 3.12.15 or newer, up to any 3.14.x (3.15 is not supported)
Big thanks to Aviationist101 for making the icon
Thanks to ilanoboy for helping me with the testing stage
You might also want to check out sblink:
https://flightsim.to/addon/108777/sblink
Or BorderlessFS:
https://flightsim.to/addon/114819/borderlessfs
Licensed under MIT License with Commons Clause v1.0. Free for personal use and modification. Selling, monetizing, or commercial distribution is strictly prohibited. See LICENSE.txt in the download for more information
Häufig gestellte Fragen
GSXComms (Ground Crew for GSX)
GSXComms is an utility designed to enhance the use of GSX in Microsoft Flight Simulator by enabling natural voice control. It processes voice commands to interact with the simulator, allowing for a streamlined experience without the need for manual navigation through the GSX panel. The application operates completely offline, ensuring privacy and efficiency, while automatically adapting the ground crews accent to match the airport location.
Nutzerbewertungen
7 Bewertungen
Versionsgeschichte
Major Update GSXCommsV4 - New speech recognition models, much higher accuraccy at much lower performance cost - Brand new speech recognition system - New launcher, with a much more beaultiful GUI - Force accent - Randomize gender - Ignore speed restriction - Simulation mode (doesn't require SimConnect or GSX Connection) - Bug fixes - Performance improvements - More Have fun!!
Ähnliche Addons

BorderlessFS
BorderlessFS is an application that allows users to run Microsoft Flight Simulator 2024 in Borderless Windowed, Fullscreen, or Windowed modes. It is designed to simplify mode selection, especially since Borderless Windowed is not natively available in MSFS 2024 settings. The tool ensures smoother interaction with other applications while maintaining in-game performance. Distributed under the MIT with Commons Clause license, modification is allowed but commercial use is prohibited.
von ziul510

sblink
sblink is an external application that manages Microsoft Flight Simulator addons by linking only those required for your current SimBrief flight plan to the Community folder. It helps reduce performance loss caused by numerous inactive addons. Key features include automatic addon linking, addon mapping, and background operation. Requires Python 3.14 and is not intended to be used alongside MSFS Addon Linker.
von ziul1010

Melden Sie sich an, um an der Diskussion teilzunehmen
AnmeldenKommentare(149)
about 2 months ago
about 3 hours ago
It still only activates GSX deboarding when I ask for boarding, even though GSXComms says it has activated boarding, so it definitely heard me correctly.
Also I get this error when asking for a list of menu items:
[CMD]: GSX menu readout requested
[GSX]: Reading 9 GSX menu entries aloud
[ERR]: GSX menu readout failed: Kokoro could not initialize OpenPhonemizer: Weights only load failed. This file can still be loaded, to do so you have two options, [1mdo those steps only if you trust the source of the checkpoint[0m.
(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
WeightsUnpickler error: Unsupported global: GLOBAL dp.preprocessing.text.Preprocessor was not an allowed global by default. Please use `torch.serialization.add_safe_globals([dp.preprocessing.text.Preprocessor])` or the `torch.serialization.safe_globals([dp.preprocessing.text.Preprocessor])` context manager to allowlist this global if you trust this class/function.
Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.
about 2 hours ago
Can you please post the console output?
The one in the main app ui, the main one, not the debug menu one
11 days ago
It seems to be working now with the Mistral option. I am not sure what I have done to make it work but I went back in to check that the API key was correctly entered, it was, but I clicked on the save button in settings anyway. It is possible that I did not click the save button when selecting Mistral the first time I used it.
Another issue is that when I request boarding of passengers it always activates de-boarding instead. The acknowledgment from GSXComms is correct for both boarding and de-boarding so it appears to recognises my voice commands correctly but activates the wrong action for boarding.
11 days ago
That's really strange
I’ll tighten the code so that when it hears “boarding,” it triggers only boarding, and when it hears “deboarding,” it triggers only deboarding
:)
10 days ago
10 days ago
Dieser Kommentar wurde gelöscht
12 days ago
I would like to ask how do you ask for a gate on arrival and how or when you ask for pushback direction. Thanks
11 days ago
Gate selection is all over the place depending on the airport, and I couldnt find an easy way that works for every airport (or even most). So for this one, you’ll have to use the menu navigation feature
You can turn it on in settings. If you want, you can even have the tts read it out so you can pick the option. Oh, and make sure Broader Mode is on, because gate names can be kinda weird (like A6 shows up as “A 6” ingame)
12 days ago
----------------------------------------------------------------------------------------------------------------
Unfortunately not working for me.
I have Python v1.14.3 installed.
I successfully downloaded the extra voice recognition files and got the required API key as per the instructions in the settings.
I have tried all three voice recognition settings options.
In all cases when I press the M key I hear the static noise.
No response to my commands.
GSXComms shows the following error:
[ERR]: Whisper transcription failed: Library cublas64_12.dll is not found or cannot be loaded
Internet search says that cublas64_12 is an Nvidia dll. (I don't know what that has to do with voice recognition).
I am using a current version of Nvidia GPU driver.
Running nvidia-smi in a command window shows that my CUDA version is v13.3
I have done a complete search of my system drive and there are no files that begin with cublas on my system, so cublas64_12.dll does not exist on my PC.
12 days ago
First of all, thank you for taking some time and doing some research to try fixing it
This is a known issue with the current version. The next update will include a fully reworked voice recognition system and some other things.
For now, the best option would be to use the Mistral API. I see you mentioned it didnt, which is strange, because this errors occurs with the whisper models. Please try putting your key there, then fully restarting your PC, if that doesn't work, you might want to manually edit the config Json (and I forgot the name of the file right now) to enable it.
If you need any help, send me a private message :)
14 days ago
14 days ago
:)
14 days ago
14 days ago
The app has a male and female response per accent for each "main" service (the ones you can request at the gate)
This includes "Cockpit to Ground"
However, other services/gsx menu navigation will not get you a realistic responde
:)
16 days ago
[ERR]: Whisper transcription failed: Library cublas64_12.dll is not found or cannot be loaded
What did i wrong ?
16 days ago
For now, you can use the medium whisper model or the Mistral provider (faster and better quality, recommended)
16 days ago
I use the mid model but the recognize of my spoken commands is very poor, so I hope the min Model works better.
At last I try to create an Mistral API key .
16 days ago
I am looking forward to upgrading the model
If you need any help creating the API key, just send me a message and I can help you with it ;)
15 days ago