Problems Getting Meshages TTS to work #130

New Issue

2026-10-05T01:25:07+02:00

adam commented

2026-10-05 01:25:07 +02:00

Hi, thanks for all the amazing work you have done with this project!

I'm trying to get Meshages TTS to work. I have followed the radio.md instructions. I have tested KittenTTS with a simple script, it was able to generate an output.wav which I used aplay to play on the default sound device. That works, I could hear the speech. I have restarted mesh_bot, and even tried rebooting the computer running mesh_bot.

Maybe I don't understand what this section of radioMon does (below). What I'm expecting is that when a transmission comes into channel 1 (because of ttsChannels = 1), I will hear it read out loud on my computer (Ubuntu 22.04) by the magic of KittenTTS. I'm not intending to use VOX, just TTS.

These are the settings from config.ini that I am using. I haven't fully investigated what most of this does — I'm a beginner — please forgive me any mistakes here.

Thanks for your help!

[radioMon]
# dx cluster `dx` command
dxspotter_enabled = True

# alerts in this module use the following interface and channel
sigWatchBroadcastInterface = 1
# broadcast channel can also be a comma separated list of channels
sigWatchBroadcastCh = 1

# using Hamlib rig control will monitor and alert on channel use
enabled = False
rigControlServerAddress = 127.0.0.1:4532
# minimum SNR as reported by radio via hamlib
signalDetectionThreshold = -10
# hold time for high SNR
signalHoldTime = 10
# the following are combined to reset the monitor
signalCooldown = 5
signalCycleLimit = 5

# Enable VOX detection using default input
voxDetectionEnabled = False
# description to use in the alert message
voxDescription = VOX

useLocalVoxModel = False
# default language for VOX detection
voxLanguage = en-us
# sound.card input device to use for VOX detection, 'default' uses system default
voxInputDevice = default
# "hey chirpy"
voxOnTrapList = False
voxTrapList = chirpy
# allow use of 'weather' and 'joke' commands via VOX
voxEnableCmd = False

# Meshages Text-to-Speech (TTS) for incoming messages and DM
meshagesTTS = True
ttsChannels = 1

# WSJT-X UDP monitoring - listens for decode messages from WSJT-X, FT8/FT4/WSPR etc.
wsjtxDetectionEnabled = False
# UDP address and port where WSJT-X broadcasts (default: 127.0.0.1:2237)
wsjtxUdpServerAddress = 127.0.0.1:2237
# Comma-separated list of callsigns to watch (empty = all callsigns)
wsjtxWatchedCallsigns =

# JS8Call TCP monitoring - connects to JS8Call API for message forwarding
js8callDetectionEnabled = False
# TCP address and port where JS8Call API listens (default: 127.0.0.1:2442)
js8callServerAddress = 127.0.0.1:2442
# Comma-separated list of callsigns to watch (empty = all callsigns)
js8callWatchedCallsigns =

Hi, thanks for all the amazing work you have done with this project! I'm trying to get Meshages TTS to work. I have followed the radio.md instructions. I have tested KittenTTS with a simple script, it was able to generate an output.wav which I used aplay to play on the default sound device. That works, I could hear the speech. I have restarted mesh_bot, and even tried rebooting the computer running mesh_bot. Maybe I don't understand what this section of radioMon does (below). What I'm expecting is that when a transmission comes into channel 1 (because of `ttsChannels = 1`), I will hear it read out loud on my computer (Ubuntu 22.04) by the magic of KittenTTS. I'm not intending to use VOX, just TTS. These are the settings from config.ini that I am using. I haven't fully investigated what most of this does — I'm a beginner — please forgive me any mistakes here. Thanks for your help! ``` [radioMon] # dx cluster `dx` command dxspotter_enabled = True # alerts in this module use the following interface and channel sigWatchBroadcastInterface = 1 # broadcast channel can also be a comma separated list of channels sigWatchBroadcastCh = 1 # using Hamlib rig control will monitor and alert on channel use enabled = False rigControlServerAddress = 127.0.0.1:4532 # minimum SNR as reported by radio via hamlib signalDetectionThreshold = -10 # hold time for high SNR signalHoldTime = 10 # the following are combined to reset the monitor signalCooldown = 5 signalCycleLimit = 5 # Enable VOX detection using default input voxDetectionEnabled = False # description to use in the alert message voxDescription = VOX useLocalVoxModel = False # default language for VOX detection voxLanguage = en-us # sound.card input device to use for VOX detection, 'default' uses system default voxInputDevice = default # "hey chirpy" voxOnTrapList = False voxTrapList = chirpy # allow use of 'weather' and 'joke' commands via VOX voxEnableCmd = False # Meshages Text-to-Speech (TTS) for incoming messages and DM meshagesTTS = True ttsChannels = 1 # WSJT-X UDP monitoring - listens for decode messages from WSJT-X, FT8/FT4/WSPR etc. wsjtxDetectionEnabled = False # UDP address and port where WSJT-X broadcasts (default: 127.0.0.1:2237) wsjtxUdpServerAddress = 127.0.0.1:2237 # Comma-separated list of callsigns to watch (empty = all callsigns) wsjtxWatchedCallsigns = # JS8Call TCP monitoring - connects to JS8Call API for message forwarding js8callDetectionEnabled = False # TCP address and port where JS8Call API listens (default: 127.0.0.1:2442) js8callServerAddress = 127.0.0.1:2442 # Comma-separated list of callsigns to watch (empty = all callsigns) js8callWatchedCallsigns = ```

adam closed this issue

2026-10-05 01:25:07 +02:00

adam commented

2026-10-05 01:25:08 +02:00

I saw you had a new release, so I tried update.sh and while that seemed to work, it didn't seem to fix the issue.

I saw you had a new release, so I tried update.sh and while that seemed to work, it didn't seem to fix the issue.

adam commented

2026-10-05 01:25:09 +02:00

is this any help? https://github.com/SpudGunMan/meshing-around/blob/main/modules/radio.md

is this any help? https://github.com/SpudGunMan/meshing-around/blob/main/modules/radio.md

adam commented

2026-10-05 01:25:10 +02:00

I followed the instructions there originally. And it didn't work.

I don't know python very well yet, so I apologize if I just missed something.

I think the issue is when you do this:

pip install https://github.com/KittenML/KittenTTS/releases/download/0.1/kittentts-0.1.0-py3-none-any.whl

It doesn't get into the virtual environment that the meshbot user is running in. It is put in the account you login as.

I altered the launch.sh to add a home directory:

export HOME=/opt/meshing-around

Then I made a .cache file:

$ ls -ld /opt/meshing-around/.cache
drwxr-xr-x 4 meshbot meshbot 4096 Apr 14 18:13 /opt/meshing-around/.cache

Then I ran:

sudo -u meshbot bash -c 'export HOME=/opt/meshing-around; source /opt/meshing-around/venv/bin/activate; pip install https://github.com/KittenML/KittenTTS/releases/download/0.1/kittentts-0.1.0-py3-none-any.whl'

That got me to this log message, which it read out loud:
System: RadioMon: Generating TTS for text: Hey its Cheerpy! Thanks for using Meshing-Around on Meshtasstic! with voice: expr-voice-2-m

When I set ttsChannels = 1 and send a message to Channel 1, it says:

INFO | Device:1 Channel:1 Ignoring Message: Hello world From: MYNODE

And doesn't read it out.

I think that is because all channels except for 0, the public channel are set to be monitored for commands, so it "ignores" non commands. Maybe that is why it isn't reading the channel outloud? (Though it doesn't read off any command there either, like when I run ping).

Any clues what I might be doing incorrectly? Was there some other way to install the kittentts python package that I didn't understand?

Thanks again!

I followed the instructions there originally. And it didn't work. **I don't know python very well yet, so I apologize if I just missed something.** I think the issue is when you do this: `pip install https://github.com/KittenML/KittenTTS/releases/download/0.1/kittentts-0.1.0-py3-none-any.whl` It doesn't get into the virtual environment that the meshbot user is running in. It is put in the account you login as. I altered the launch.sh to add a home directory: `export HOME=/opt/meshing-around` Then I made a .cache file: ``` $ ls -ld /opt/meshing-around/.cache drwxr-xr-x 4 meshbot meshbot 4096 Apr 14 18:13 /opt/meshing-around/.cache ``` Then I ran: ``` sudo -u meshbot bash -c 'export HOME=/opt/meshing-around; source /opt/meshing-around/venv/bin/activate; pip install https://github.com/KittenML/KittenTTS/releases/download/0.1/kittentts-0.1.0-py3-none-any.whl' ``` That got me to this log message, which it read out loud: `System: RadioMon: Generating TTS for text: Hey its Cheerpy! Thanks for using Meshing-Around on Meshtasstic! with voice: expr-voice-2-m` When I set `ttsChannels = 1` and send a message to Channel 1, it says: `INFO | Device:1 Channel:1 Ignoring Message: Hello world From: MYNODE` And doesn't read it out. I think that is because all channels except for 0, the public channel are set to be monitored for commands, so it "ignores" non commands. Maybe that is why it isn't reading the channel outloud? (Though it doesn't read off any command there either, like when I run ping). Any clues what I might be doing incorrectly? Was there some other way to install the kittentts python package that I didn't understand? Thanks again!

adam commented

2026-10-05 01:25:10 +02:00

sorry for delay, do you have ignore channel on somewhere? is there a public channel ignore set?

sorry for delay, do you have ignore channel on somewhere? is there a public channel ignore set?

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: adam/meshing-around#130