#meshing-around BBS/utility BOT
1 messages · Page 7 of 1
Yeah, every time I do this, I learn a little bit more about things.
It’s a combination sometimes doing the same thing more than once. Or realizing the problem is something completely different and you’re just seeing the result of it. I must’ve tried every variation on the config file. And yet it the node was hung on another app and a few other things I won’t go into length about. Mahalo 🥔 gunman
Aloha- I'm getting an error on this url- yet when i try it - works... any thoughts people
`# NOAA Coastal Data Enable NOAA Coastal Waters Forecasts and Tide
coastalEnabled = True
Find the correct costal weather directory at https://tgftp.nws.noaa.gov/data/forecasts/marine/coastal/
pz = Puget Sound, ph = Honolulu HI, gm = Florida Keys, pk = Alaska
this map can help https://www.weather.gov/marine select location and then look at the 'Forecast-by-Zone Map'
myCoastalZone is the .txt file with the forecast data
myCoastalZone = https://tgftp.nws.noaa.gov/data/forecasts/marine/coastal/hi/hiz026.txt
number of data points to return, default is 3
coastalForecastDays = 3`
spud, have you interfaced with i2c at all from mesh_bot?
I was thinking about seeing if I could have it include current 12V and 5V current and voltage measurements in the sysinfo
or sitrep
or something
... specifically on a FemtoFox, in my case
You want a block code section for that kind of copypasta. It uses 3x ` (grave/backtick) at the start and end and looks like this:
#look at this comment
lines: 3
a_third_line: true
You can put c or cpp at the front, too
Doh
Oh, sure. I was using yaml as the language spec in my example. I should've mentioned that.
Discord supports a bunch, like py, js, cs, and probably some others that I don't use 😛
There's a news text file that it monitors and broadcasts
so inset this at the bottom - `https://tgftp.nws.noaa.gov/data/forecasts/marine/coastal/hi/hiz026.txt
number of data points to return, default is 3
coastalForecastDays = 3
#look at this comment
lines: 3
a_third_line: true`
Hard part would be if the Meshtasticd is reading the i2c itself, and something else tries to as well.
The perfect version of this would have meshtasticd reading out the 12V voltage and transmitting battery state like it's supposed to, but that seems ambitiously hopeful
My hope is to run 5V and 12V through the INA3221 to see why the FemtoFox keeps boot cycling
Low voltage is my best guess, hence these
But a 20W panel should be plenty for it.
In any case, it sounds like I don't need to involve mesh_bot. I assumed meshtastic doing it would be harder.
Ina3221 has 3 channels 🙂
Idk what battery you're using, but that can either be scaled in, or just read battery % ignored
not all products are alive any more this specific example is COASTAL HAZARD MESSAGE NATIONAL WEATHER SERVICE HONOLULU HI 337 PM HST TUE MAR 1 2016
so its reporting no data
the bot will check date to validate record a little before sending it over the air
just patched a few things and added a alert to the sentry channel (default is node 1 channel 2) for a low battery alarm 25% will go out as a text message
Look at you with your functional battery management
My next major to do is have any idea that batteries exist on two of my big nodes
Currently they just think they're on mains power all the time
Most of my stuff is, I didn’t test the three meter power but if node telemetry indicates primary is low (25%) it will send a 🚨( it should also prior have been in the log)
Battery chemistry needs to be a yaml defined thing
or maybe the % curve needs adding
found a working link. ph- instead of hi- up'd the output to 10 days-- good stuff.
https://tgftp.nws.noaa.gov/data/forecasts/marine/coastal/ph/phz122.txt
I'm getting this error starting the browser...
sudo journalctl -u mesh_bot.service
System: LLM failure: HTTPConnectionPool(host='172.27.0.19', port=11434): Max retries exceeded with url: /api/generate (Caused by Ne
if it's localhost instead of the ip address I get -
System: LLM failure: HTTP Error: 404
do i need this?
[Service] Environment="OLLAMA_HOST=0.0.0.0:11434"
in /etc/systemd/system/ollama.service?
Yes it will enable the ollama HTTP API
Then i did this
sudo nano /etc/systemd/system/ollama.service
`[Unit]
Description=Ollama Service
After=network-online.target
[Service]
ExecStart=/usr/local/bin/ollama serve
User=ollama
Group=ollama
Restart=always
RestartSec=3
Environment="PATH=/opt/meshing-around/venv/bin:/home/wh6gxz/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/l>
[Service]
Environment="OLLAMA_HOST=0.0.0.0:11434"
[Install]
WantedBy=default.target`
Move all service items under one heading if it didn’t work
Here's what I did...
/etc/systemd/system/mesh_bot.service:23: Failed to parse service type, ignoring:
[Service] ExecStart=/usr/local/bin/ollama serve User=ollama Group=ollama Restart=always RestartSec=3 Environment="PATH=/opt/meshing-around/venv/bin:/home/wh6gxz/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/l> Environment="OLLAMA_HOST=0.0.0.0:11434"
info:
For each environment variable, add a line Environment under section [Service]:
[Service]
Environment="OLLAMA_HOST=0.0.0.0:11434"
yeah hit recycle - and
WARNING | System: LLM failure: HTTP Error: 404
doooh...
...at lest the error Failed to parse service type, ignoring: went away on recycle...
So good now?
WARNING | System: LLM failure: HTTP Error: 404
doooh...
This from meshbot make sure the url is accessible (try curl or browser should get ollama ready)
Device:1 Channel:1 ReceivedChannel: askai kiaua volcano From: w> Aug 28 12:03:23 VolcanoAI bash[1923]: 2025-08-28 12:03:23,660 | INFO | Device:1 Channel:1 SendingChannel: Please wait, response could > Aug 28 12:03:25 VolcanoAI bash[1923]: 2025-08-28 12:03:25,864 | WARNING | System: LLM failure: HTTP Error: 404 Aug 28 12:03:28 VolcanoAI bash[1923]: 2025-08-28 12:03:28,064 | INFO | Device:1 Channel:1 SendingChannel: ⛔️I am having trouble proces
I restart ollama and i get this output
/etc/systemd/system/mesh_bot.service:23: Failed to parse service type, ignoring: notify #try simple
🙃
config the file this way parsing error is gone...
`Service]
ExecStart=/usr/local/bin/ollama serve
User=ollama
Group=ollama
Restart=always
RestartSec=3
Environment="PATH=/opt/meshing-around/venv/bin:/home/wh6gxz/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/l>
[Service]
Environment="OLLAMA_HOST=0.0.0.0:11434"`
so i kept the same as above ran the url... and ok then
...and
Device:1 Channel:1 ReceivedChannel: askai kiaua volcano From: w> Aug 28 12:24:28 VolcanoAI bash[1023]: 2025-08-28 12:24:28,602 | INFO | Device:1 Channel:1 SendingChannel: Please wait, response could > Aug 28 12:24:30 VolcanoAI bash[1023]: 2025-08-28 12:24:30,806 | WARNING | System: LLM failure: HTTP Error: 404 Aug 28 12:24:33 VolcanoAI bash[1023]: 2025-08-28 12:24:33,006 | INFO | Device:1 Channel:1 SendingChannel: ⛔️I am having trouble proc
ok... step back - i'm going circular
2025-08-28 12:34:21,188 | WARNING | System: LLM failure: HTTP Error: 404 Aug 28 12:34:23 VolcanoAI bash[1023]: 2025-08-28 12:34:23,388 | INFO | Device:1 Channel:1 SendingChannel: ⛔️I am having trouble proces> Aug 28 12:34:49 VolcanoAI systemd[1]: /etc/systemd/system/mesh_bot.service:23: Failed to parse service type, ignoring:
Ollama is its own service and it runs with the http web enabled when it has the environment variable in its [service] record
so it explians the url working and the bot out being other
yeah I followed that earlier.
I took out
Environment="OLLAMA_HOST=0.0.0.0:11434"
and--
System: LLM failure: HTTPConnectionPool(host='192.168.86.24', port=11434): Max retries exceeded with url: /api/generate
It will need ollama host to auto enable the server
Is there any security of the OS stopping it
So if the meshbot is running and fails on LLM HTML that’s fine just leave it for now - focus on getting ollama web api online
It should pop up in a browser
In a generic Debian OS there is no security to stop this
Check the journal on ollama service is it a bind fail?
Are the interfaces real or all hosted cloud space?
Can you use loop back if hosted?
Or set the IP to the interface IP not 0.0.0.0
yeah it try's to run on a ip choke with this...
: Aug 28 12:49:51 VolcanoAI bash[2116]: 2025-08-28 12:49:51,535 | INFO | System: Autoresponder Started for Device1 VolcanoAI,Vail. NodeI> Aug 28 12:49:51 VolcanoAI bash[2116]: 2025-08-28 12:49:51,535 | WARNING | System: These LLM models lack a traditional system prompt, they> Aug 28 12:49:51 VolcanoAI bash[2116]: 2025-08-28 12:49:51,537 | WARNING | System: LLM failure: HTTPConnectionPool(host='172.27.0.19', por> Aug 28 12:50:25 VolcanoAI systemd[1]: /etc/systemd/system/mesh_bot.service:23: Failed to parse service type, ignoring: notify #try simple > Aug 28 12:51:27 VolcanoAI systemd[1]: Stopping mesh_bot.service - MESH-BOT... Aug 28 12:51:27 VolcanoAI systemd[1]: mesh_bot.service: Deactivated successfully. Aug 28 12:51:27 VolcanoAI systemd[1]: Stopped mesh_bot.service - MESH-BOT. Aug 28 12:51:33 VolcanoAI systemd[1]: Started mesh_bot.service - MESH-BOT. Aug 28 12:51:34 VolcanoAI bash[2172]: Aug 28 12:51:34 VolcanoAI bash[2172]: Meshtastic Autoresponder Bot CTL+C to exit Aug 28 12:51:34 VolcanoAI bash[2172]: Aug 28 12:51:34 VolcanoAI bash[2172]: 2025-08-28 12:51:34,154 | INFO | System: Autoresponder Started for Device1 VolcanoAI,Vail. NodeI> Aug 28 12:51:34 VolcanoAI bash[2172]: 2025-08-28 12:51:34,154 | WARNING | System: These LLM models lack a traditional system prompt, they> Aug 28 12:51:34 VolcanoAI bash[2172]: 2025-08-28 12:51:34,156 | WARNING | System: LLM failure: HTTPConnectionPool(host='172.27.0
Yes this looks fine; just need the ollama API online
Web browse to the http://localhost:11434
When you see ollama api ready or whatever it says it’s good to cook
mahalo - I'll install the API- Here's to nurse dude
ok did all that re-inserted the path in the service file. url works despite bot output...
bot output on channel
circle of errors...
Device:1 Channel:1 ReceivedChannel: askai kiaua volcano From: wh6gxz-ecom NFO | Device:1 Channel:1 SendingChannel: Please wait, response could take 30+ seconds. Fund the SysOp's GPU budget! ING | System: LLM failure: HTTP Error: 404 NFO | Device:1 Channel:1 SendingChannel: ⛔️I am having trouble processing your request, please try again later.
Is there connectivity from the bot to the url?
Can you see ollama api in a browser
yes the bot config file has the url
INFO | Device:1 Channel:1 ReceivedChannel: askai From: wh6gxz-ecom Aug 28 13:46:22 VolcanoAI bash[1952]: 2025-08-28 13:46:22,045 | INFO | Device:1 Channel:1 SendingChannel: Please ask a question Aug 28 13:46:51 VolcanoAI bash[1952]: 2025-08-28 13:46:51,630 | INFO | Device:1 Channel:1 ReceivedChannel: askai what is the big islan> Aug 28 13:46:51 VolcanoAI bash[1952]: 2025-08-28 13:46:51,630 | INFO | Device:1 Channel:1 SendingChannel: Please wait, response could > Aug 28 13:46:53 VolcanoAI bash[1952]: 2025-08-28 13:46:53,835 | WARNING | System: LLM failure: HTTP Error: 404 Aug 28 13:46:56 VolcanoAI bash[1952]: 2025-08-28 13:46:56,035 | INFO | Device:1 Channel:1 SendingChannel: ⛔️I am having trouble proce
have ollama running - 2 models downloaded
`ollama run gemma3:270m
who is spudgunman?
Spudgunman is a fictional character from the popular video game "Minecraft". He is a powerful and cunning character who is a powerful
fighter and a skilled strategist.`
wiki: is COOL 
also:
wh6gxz@VolcanoAI:~ $ ollama list NAME ID SIZE MODIFIED deepseek-r1:1.5b e0979632db5a 1.1 GB 15 minutes ago gemma3:270m e7d36fb2c3b3 291 MB 2 days ago
haha
who is tom hanks?
Tom Hanks is a fictional character from the popular video game "Minecraft."
ohhh AI -- you LOLO!
reboot- then or - /bye - then run again can make ai happy for the time LOL
ollama run gemma3:270m
who is tom hanks?
Tom Hanks is an American actor, known for his roles in films like "A Streetcar Named Desire," "The Shawshank Redemption," and "The
Lord of the Rings: The Return of the Lads."
wiki: works better for info...
Haha you just joined in on gemma3:270
It’s the latest in Gemma code, if you want an AI with a prompt let’s say 2024 version it can use context from google and have some legitimacy ( I didn’t delete old way yet just flip off raw mode and use Gemma2)
If you have a full GPU you can scale up Gemma3:latest
Or go larger
Latest is only 4b or about 4gb ram, more ram “more smarts”
The default meshbot is the smallest model at 300mb
It’s dumb but funny why I holding there at the moment while I goof myself with newest model and I wanna try and use it possibly for more tooling or direct code interactions
I want it potato easy turnkey, like perfect hash browns it’s not ready yet
Yeah, fun stuff. I still don’t have the bot picking up AI. I am running it on the command line.
I get this output on the bot
bot log
System: LLM failure: HTTPConnectionPool(host='172.27.0.1', port=11434): Max retries exceeded with url: /api/generate (Caused by NewC> NFO | Device:1 Channel:1 SendingChannel: ⛔️I am having trouble processing your request, please try again later
journalctl -e -u ollama
[GIN] 2025/08/28 - 19:47:01 | 404 | 6.537µs | 192.168.86.42 | GET "/favicon.ico"
yeah, can't get the bot to output ai--
i get the above output once after restarting the bot then this
restart bot then output returns to
System: LLM failure: HTTPConnectionPool(host='172.27.0.1', port=11434): Max retries exceeded with url: /api/generate (Caused by NewC> NFO | Device:1 Channel:1 SendingChannel: ⛔️I am having trouble processing your request, please try again later
I'm going to try gemma2 with the bot-
same errors - play with the command line a bit-
no it fails how ever i config- it -
otherwise the two alone are good- ...
-- put some time into it... it's all good
WARNING | System: LLM failure: HTTPConnectionPool | System: LLM failure: HTTPConnectionPool(host='172.27.0.1', port=11434): Max retries exceeded with url: /api/generate (Caused by NewonnectionError('<urllib3.connection.HTTP Connection object at 0x7fffab69fe90>: Failed to establish a new connection:=11434): Max retries exceeded with url: /api/generate (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7fffab>
yes- tried the 172- and variety of url variables - have a port open in firewall 11434
I see 172?
let me try that again-- i've tried so many variables hhi
BOT output:
``WARNING | System: LLM failure: HTTP Error: 404
Aug 29 07:58:53 VolcanoAI bash[2454]: 2025-08-29 07:58:53,351 | INFO | Device:1 Channel:1 ReceivedChannel: askai hawaii From: hitchike>
Aug 29 07:58:53 VolcanoAI bash[2454]: 2025-08-29 07:58:53,988 | INFO | Device:1 Channel:1 SendingChannel: Please wait, response could >
Aug 29 07:58:56 VolcanoAI bash[2454]: 2025-08-29 07:58:56,191 | WARNING | System: LLM failure: HTTP Error: 404`
VoilcanoAI node output
`Aug 29 07:58:58 VolcanoAI bash[2454]: 2025-08-29 07:58:58,391
| INFO | Device:1 Channel:1 SendingChannel: ⛔️I am having``
ollama output...
VolcanoAI ollama[973]: [GIN] 2025/08/29 - 07:58:15 | 404 | 2.694251ms | 127.0.0.1 | POST "/api/generate" Aug 29 07:58:56 VolcanoAI ollama[973]: [GIN] 2025/08/29 - 07:58:56 | 404 | 359.224µs | 127.0.0.1 | POST "/api/generate"
command line works...
Ollama log
GIN] 2025/08/29 - 08:02:43 | 200 | 4.017066046s | 127.0.0.1 | POST "/api/generate" Aug 29 08:03:00 VolcanoAI ollama[973]: [GIN] 2025/08/29 - 08:03:00 | 200 | 1.151286535s | 127.0.0.1 | POST "/api/chat" Aug 29 08:04:52 VolcanoAI ollama[973]: [GIN] 2025/08/29 - 08:04:52 | 200 | 40.204µs | 127.0.0.1 | HEAD "/" Aug 29 08:04:52 VolcanoAI ollama[973]: [GIN] 2025/08/29 - 08:04:52 | 200 | 35.278µs | 127.0.0.1 | GET "/api/ps"
Gotta be close, or there is specifically security stopping it
Try a curl?
Is it the meshbot user
No access
something is making making this happen
`System: LLM failure: HTTP Error: 404``
The inability to access via http
if i put the real ip - ollama will output on a browser
I tried curl- how ru suggesting using it?
wh6gxz@VolcanoAI:~ $ curl http://localhost:11434/
Ollama is running
Just gotta solve it for the meshbot service now
If you stop meshbot
And run it from command line it is failing to find ollama
no - it outputs...
curl http://localhost:11434/
Ollama is running
Yea that’s good so the bot can’t seem to do the same thing
I suspect permissions but it’s finding what they are
The specific code is if a return of 200 use data any other error with http alert to screen
The 404 is either a typo or something causing a hiccup
(It’s outside my ability to fix localized issue)
yeah - the bot is not loading LLM api HTML fails... - mahalo for working with me... otherwise - I've been playing with a few models on the pi5 16gb...
Fun! Worst case you can try a new pi image
I just added a line to install.sh to automate ollama more the other day
I will test a new install this evening if I can get time
I've learned alot BRO! Having fun... I'll fool around with it on and off until - well - we will see about that- 73 and aloha - I'll be your tester hihi love the BOT!
this is running behind a arden hap- the gateway address is 404
journalctl -e -u ollama
2025/08/29 - 09:18:51 | 404 | 4.171185ms | 172.27.0.1 | POST "/api/generate" Aug 29 09:18:51 VolcanoAI ollama[973]: [GIN] 2025/08/29 - 09:18:51 | 200 | 290.185µs | 192.168.86.42 | GET "/" Aug 29 09:19:07 VolcanoAI ollama[973]: [GIN] 2025/08/29 - 09:19:07 | 404 | 381.111µs | 172.27.0.1 | P
i think this is why i'm foof-
journalctl -e -u ollama
[GIN] 2025/08/29 - 09:51:42 | 200 | 8.612µs | 192.168.86.42 | GET "/" Aug 29 09:52:15 VolcanoAI ollama[978]: [GIN] 2025/08/29 - 09:52:15 | 404 | 377.276µs | 172.27.0.19 | POST "/api/generate"
if i do the 192.x. in the config file - I get a 172.27.0.1 gateway ip- 404-
404 isn’t a valid IP address but it is a response in failure
If this is all on one raspberry pi I’m not sure how it got so tangled up
Try a reboot haha
yeah done a few of those hihi.
i moved it to another part of my network just to make sure routing was not the issue...
pretty much reproduced the same stuff. 🤪 💯 👀
Is everything on one pi? The bot and ollama?
pi5 16gb -- everything else is cool 🙃 ...
ollama run gemma3:270m
who is spudgunman
Spudgunman is a fictional character in the video game "The Legend of Zelda: Breath of the Wild." He is a character from the game.
`# Enable ollama LLM see more at https://ollama.com
ollama = True
Ollama model to use (defaults to gemma3:270m)
ollamaModel = gemma3:latest
server instance to use (defaults to local machine install)
ollamaHostName = http://localhost:11434
Produce LLM replies to messages that aren't commands?
If False, the LLM only replies to the "ask:" and "askai" commands.
llmReplyToNonCommands = True
if True, the input is sent raw to the LLM, if False uses legacy template query
rawLLMQuery = True
`
config.ini
Looks good
So what are the requirements for a Ollama server ?
Like if I was to spin up a dedicated VM to just be a Ollama server for this project..
The new default is gemma3 @ 300mb ram 🐏 the more the cpu the faster it is and if you have a GPU it blazes
I dev the bot on a pi5 4GB
Lab is pi5 8gb
And my production is a gpu system
You can put ollama on gpu and low power bot
HTTP call
I have a colo in a datacenter, with virtual systems running. So I could give it like 4 CPUs, 16GB RAM, and 100GB HDD w/o any issues..
Ollama is fun if you never goofed with it
I've never heard of it till this project..
I haven't heard of Ollama as well before meshing-around... 
Other than not getting ollama output on the bot it's all good-
i've have meshing-around on a pi4, femtofox, pi5 - oh yeah
That's kinda my plan.. I'm trying to build out the mesh in my area, I am completely cut off from all meshes..
VolcanoAI 12:45 PMAug 29, 2025, 12:45 PM MOTD: Thanks for using Nurse Dude BOT! Mesh ON!

i should say mahalo--
VolcanoAI 12:47 PMAug 29, 2025, 12:47 PM MOTD: Mahalo for using Nurse Dude BOT! Mesh ON!
haha there ya GO brah!
askai
Why is wxalert listed twice at the end ?
yeah- just noticed that myself...
ran cmd again and yep it be 2 hihi
Just noticed it myself
I know why
Shouldn’t be that way
Also looks like you got it working? @heady raven
No unfortunately
cmd askai hawaii
BOT LOG
System: LLM failure: HTTP Error: 404 NFO | Device:1 Channel:1 ReceivedChannel: askai hawaii From: hitchiker-udp NFO | Device:1 Channel:1 SendingChannel: Please wait, response could take 30+ seconds. Fund the SysOp's GPU budget! ING | System: LLM failure: HTTP Error: 404 NFO | Device:1 Channel:1 SendingChannel: ⛔️I am having trouble processing your request, please try again later.
Ollama LOG
Aug 29 13:07:41 VolcanoAI ollama[1359]: [GIN] 2025/08/29 - 13:07:41 | 404 | 379.021µs | ::1 | POST "/api/generate" Aug 29 13:08:44 VolcanoAI ollama[1359]: [GIN] 2025/08/29 - 13:08:44 | 404 | 357.298µs | ::1 | POST "/api/generate"
yeah if I trigger the askai without a question I get output -
Oh shoot the LLM says something like that as well
I will try a new install this evening
I haven’t had any issues with it lately like this
this is ollama via- cmd line log output... got to like 200..
ollama[1359]: [GIN] 2025/08/29 - 13:13:55 | 200 | 3.987420186s | 127.0.0.1 | POST "/api/generate" Aug 29 13:14:07 VolcanoAI ollama[1359]: [GIN] 2025/08/29 - 13:14:07 | 200 | 1.684952089s | 127.0.0.1 | POST "/api/chat" Aug 29 13:15:09 VolcanoAI ollama[1359]: [GIN] 2025/08/29 - 13:15:09 | 200 | 1.651722282s | 127.0.0.1 | POST "/api/chat" Aug 29 13:15:35 VolcanoAI ollama[1359]: [GIN] 2025/08/29 - 13:15:35 | 200 | 2.095934593s | 127.0.0.1 | POST "/api/chat"
I'm happy to recreate the mesh-wheel hihi
I did just grab a new pi SD card and it everything worked in a new build. Installing with moving to /opt no to embedded yes to mesh service and meshbot account creation yes yes yes and reboots into a working bot
@gentle kestrel What kinda GPU you using ? I have an old 1050Ti I might be able to use..
2 actually.. LOL
They work great
Those would have to be a home setup tho..
Can both be used at the same time for the 1 Ollama server ?
Yes ollama should detect them
Interesting..
Imma have to dig them out and find a desktop setup to put them in..
Oh.. I think I have a 1070Ti also.. I don't remember what I have in my other older desktops I have laying around..
If you go that far OpenWebUI is another cool project to look at with ollama stacking models and images
All uses the API so can all use the one machine
Question - do I use the scheduler module to facilitate bbslinking between nodes? I'm struggling to get things connected and maybe I'm just not quite grasping it.
I've got two nodes with a shared channel, I've got the node IDs whitelisted on each, but can't seem to get a post on one to propagate to the other.
yeah - still can't get ai to output to the bot-
[GIN] 2025/08/29 - 15:54:14 | 404 | 8.087556ms | ::1 | POST "/api/generate" Aug 29 15:56:56 VolcanoAI ollama[1115]: [GIN] 2025/08/29 - 15:56:56 | 404 | 373.417µs | ::1 | POST "/api/generate"
Correct using the schedule to set up a synch like 2am or something the message having the keyword like
bbslink MeshBot looking for peers
You can put that in config.ini vs the current hello
Thanks. I'll give it a go
So you formatted a new pi and can’t make it work
I tried it like a hour ago and it was fine
Worked perfect loaded the bot and no errors on LLM load
If you say yes to the ollama API it will also now set config.ini truly turn key on a pi
I didn't reformat the pi- re-installed meshing-around. - the api installed runs. something in path-- got to break away for a bit- mahalo for staying with me... yeah i'm reproducing the same errors hihi
Yea no worries; something is funky in the pi there do you have a spare SD card you can flash new test and see if you can locate the difference?
yeah i do have a spar-- the thing is everything else works. ai doesn't load on the bot - bummer
There is some likely typo or security issue at play
yeah - true- i moved to another part of the network- nada hmmm
If it’s all on a single pi it won’t matter
I really don’t know what could have happened here it looks like you got everything set up
yeah it is- with meshadv hat--
Nice it should all work, a pi is a pi
Doooh! During
git clone curl -fsSL https://ollama.com/install.sh | sh
`curl -fsSL https://ollama.com/install.sh | sh
Cleaning up old version at /usr/local/lib/ollama
Installing ollama to /usr/local
Downloading Linux arm64 bundle
########################## 36.7%curl: (92) HTTP/2 stream 1 was not closed cleanly: PROTOCOL_ERROR (err 1)
gzip: stdin: unexpected end of file
tar: Unexpected EOF in archive
tar: Unexpected EOF in archive
tar: Error is not recoverable: exiting now`Played with this and that configuration-
It took a few times to get Ollama downloaded configured
Didn't real do anything DIFF- did the whole thing over - got better faster - geekier @ the install haha
This Meshtasticd installed better than last one- I had to use the CLI to config the node- and it kept going back to default. So yeah- happy happyDude Bot has AI aloha only on VolcanoAI channel...
Mahalo SpudGunMan!
You ROCK
wh6gxz mini mesh
11:34 PMAug 29, 2025, 11:34 PM
Askai hawaii
sweet- piled up the cmd's while the ai did it's things all the output worked!
Wahoo!
Nice work Shawn next you could build a document upload and Hawai specific training via http for offline use than when someone at a school, shelter or site with no internet access could have detailed information on events or forms the AI could answer questions about may take a bit of OCR work and training
Anyone noticed in the last week or so rlist has started to repsond with "No Data for your Region"?
Had a guy on our local mesh mention it. Sure enough when I tried it noticed the same response from several nodes making the request.
All else seems fine. Wasn't sure if it was just me or there might be a known issue. Was going to look into maybe today or tomorrow if I get time.
Could be cloudflair preventing stuff - can you browse to https://www.repeaterbook.com from the bot?
I'll give it a try when I can get back to it.
If you get hit with a captive portal from the hosting provider cloudflair
I’m curious if it works fine after for a bit
Wasn't sure if maybe repeaterbook may have made a change as well.
Possibly; it’s not an official api I’m scraping
Been consistent now for several days apparently. Just haven't had time to look at it much. Thanks for the suggestion.
Yea thanks for alert on it, I will look more into it this evening
Hey that would be awesome. Could just be us.
I'm no python god but I had started to glace though the code and was getting an idea of how it was making calls to repeaterbook.
Using BeautifulSoup to parse the html
It's been working amazing for our small mesh running on a FEMTO. Really appreciate all the work that has been put into it.
Awesome, Glad it’s fun!
This is for sure looking like a webAppFirewall which is looking for human
I may need to retire this command 🙁 if I can’t find some good data elsewhere artsci has much less
Darn. I was somewhat afraid of that.
More digging hope isn’t lost; looks like they might have changed the website
It's a whirlwind of emotion! 😂
Try git pull
Device:1 Channel:1 ReceivedChannel: rlist From: KB0SDY SCAP2
2025-08-31 20:57:28,065 | DEBUG | System: Bot detected Commands:[{'cmd': 'rlist', 'index': 0}] From: KB0SDY SCAP2
2025-08-31 20:57:28,065 | DEBUG | System: location data for 3297246898 is 34.99878,-86.7078466
2025-08-31 20:57:30,610 | INFO | Device:1 Sending DM: KB4CRG📶442.775+5.0 MHz,107.2.FM KB4CRG📶444.575+5.0 MHz,100.0.FM EchoLink IRLP W4BV📶147.03+0.6 MHz,114.8.FM W4LCS📶442.3+5.0 MHz,127.3.FM To: KB0SDY SCAP2
That looks like a go.
Woo-Hoo!
Nice!!
Thanks dude
Really appreciate the fix.
We have quite a collection of ham guys deploying our local mesh at lincolnmesh.us and they love to show off stuff lke rlist. LOL
BBS had a minor change to give the last 4 of the hexID who posted the message (so you can use your mobile to search)
does the bot/system do DM to email or txt message?
Yes you can set up a email/sms alert system the bbsmail is just RF node
Thanks I may try and get an image up running how does it run in a VM
As long as it has access to a meshtastic node somewhere it will be fine, I may have some time soon gonna rework docker into a better configuration but I know it’s been deployed in lots of methods
Was not sure with the New UDP stuff I was thinking the VM and bot could simulate a node and send it to other nodes on the network by UDP than out to the RF land through a base station.
Just install meshtastic for your distro here https://meshtastic.org/docs/hardware/devices/linux-native-hardware/#installation
Then, use cli or Android client to configure it's region and anything else you need like channels, name and linking. After that, you can run the bot and It will be a "radioless" node that is able to communicate with other nodes via UDP or MQTT
I wanna bundle the docker this way
Thanks will give that a try for now
I made a tiny BBS to travel with, but I think the wago connectors I used to connect the battery have slipped, and it's too #notabomb for me to fix on the train.
So here I sit, bereft of dad jokes.
Oh, I see what happend.
I unplugged my laptop and it drained the battery.
Oops
Easily fixed.
Oop
Cool
I've got an interesting one maybe.
I've running meshing-around without issue and then I started getting these errors.
System: communicating with interface1, trying to reconnect: Timed out waiting for an acknowledgment
this config.ini has been working for me - and then...
`[interface]
type = tcp
port = /dev/ttyAMA10
port = /dev/ttyUSB0
hostname = localhost
mac = 00:11:22:33:44:55`
This runs on a pi5 16gb - with a MeshAdv hat - good att- until recently it's been all good - BTW it run's meshtasticd alpha- without issue - Mahalo for reading this...
Is it reconnecting fine?
yes and no - i play with the config file do a few reboots - then wola - time passes then the error reappears. hmmm
ERROR | System: communicating with interface1, trying to reconnect: Timed out waiting for an acknowledgment
not working. yet meshtasticd sees the hat-
Yea it’s tcp I have seen it have connection issues but it should just reconnect and go on
It’s somewhat expected with tcp and ble
yeah i've tried via- tcp/ip address and a few other variables
Sure but it’s reconnecting fine?
no - not anymore
What’s the full log for failure?
You just show my one line eventually the API will crash why is it crashing
turned on debug---
the config for femtofox- currently offline because i'm running meshing-around on another device
`[interface]
type = tcp
port = /dev/ttyACM0
port = /dev/ttyUSB0
hostname = localhost`
no issues....
I just see it’s last message of retrying and that is it
Did it all hang up or did it reconnect
I can’t replicate this, did a update and all
no - i'm a little stuck- i'm thinking the hat- yet I moved it to a diff- part of the network- got the same errors. then tried hostname 0.0.0.0 it worked- moved it back to the part of the network and bummer...
INFO | System: Autoresponder Started for Device1 VolcanoAI,Vail. NodeID: 2651026246, !9e037346 ING | System: These LLM models lack a traditional system prompt, they can be verbose and not very helpful be advised. ERROR | System: communicating with interface1, trying to reconnect: Timed out waiting for an acknowledgment
0.0.0.0 won’t be good for things, use a 127.0.0.1
yeah no good- bummer buggy-ness
I switched back to the femtofox-- all good until i figure out whats up with the other node- very odd- i think it might be the hat-- more experimenting
If the bot is running on the same pi as meshtasticD 127.0.0.1 always
If you are gonna use a bot on the femtofox and ollama api on the pi5 set the ollama url in config.txt on the femtofox to the pi5 IP
(Separately you might also enjoy flipping back to Gemma2 and set LLM raw query to false. The 2024 LLM is smarter at the moment)
yeah something up with the hat- going to re-gig the pi5 for ollama and link the fox to it... see how that goes...
cuurious the femtofox doesn't show riverflow in the cmd? did something change?
If riverList = in settings is empty it won’t show up do you have something configured for a river?
I've got
# NOAA Hydrology unique identifiers, LID or USGS ID riverListDefault = 16704000
Remove the word Default there
I've got to dial down the output hihi...
so much air traffic around BI
wh6gxz-fox 7:37 AMSep 12, 2025, 7:37 AM 🚀 High Altitude Detected! NodeID:3276675423 Alt:8,514ft/2,595m
Like it’s too much airplane return(I just lowered it to 3)
I was debating narrowing to altitude as well
If this is helpful let me know
I didn’t want limit it out the gate
it's aloha friday - I get this output on the femtofox bot
:rocket: High Altitude Detected! NodeID:3276675423 Alt:8,514ft/2,595m
my config.ini
for femtofox...
`# detect anyone close to the bot
SentryEnabled = True
emailSentryAlerts = False
radius in meters to detect someone close to the bot
SentryRadius = 1000
channel to send a message to when the watchdog is triggered
SentryChannel = 2
holdoff time multiplied by seconds(20) of the watchdog
SentryHoldoff = 9
list of ignored nodes numbers ex: 2813308004,4258675309
sentryIgnoreList =`
I changed the radius....
I don't see- this in my config.ini..
highFlyingAlert = True # HighFlying Node alert highFlyingAlertAltitude = 2000 # Altitude in meters to trigger the alert highflyOpenskynetwork = True # check with OpenSkyNetwork if highfly detected for aircraft
me goofing off... hihi
Add it in!
If it didn’t find any flights it won’t list them (they need to be in open sky network adsb contributors)
:rocket: High Altitude Detected! NodeID:3276675423 Alt:8,514ft/2,595m
Same node over and over could be just bad data add to ignore list
i just replace the sentry info- with this - looks clean haha
`[sentry]
SentryEnabled = True
detect anyone close to the bot
emailSentryAlerts = False
if SMTP enabled send alert to sysop email list
SentryRadius = 100
radius in meters to detect someone close to the bot
SentryChannel = 9
holdoff time multiplied by seconds(20) of the watchdog
SentryHoldoff = 2
channel to send a message to when the watchdog is triggered
sentryIgnoreList =
list of ignored nodes numbers ex: 2813308004,4258675309
highFlyingAlert = True
HighFlying Node alert
highFlyingAlertAltitude = 2000
Altitude in meters to trigger the alert
highflyOpenskynetwork = True
check with OpenSkyNetwork if highfly detected for aircraft`
I didn't blow up the config.ini so I'll see what the output brings hihi
no issue there- mahalo 🥔 I set-up an account with opensky-
I didn’t know this level of accuracy was possible wouldn’t it detect footprints at this resolution
Hawaii we get our notifications from civil defense 🤙🙃 tsunami alerts
cool stuff 🥔
is there someway to make the bot monitor and respond in more than one channel?
nvm
the node did not have the channel 🤣
well
now its seeing it but wont respond 🤣
It should be fully channel aware zero being the intended default (long fast)
A default config will ignore and treat index 0 like a public channel
okay command prefix fucks it
with it disabled it works
cmdBang = True
if false everything works
i enabled explicitcmd and it detected ping in the middle of the sentence
mhmm
just ping???
interesting
if u type test ping it respomds to it
Do a git pull I hopefully fixed now
Thanks for pointing out, fixed two bugs!
haha np
works now
now i just need to figure out meshtasticd and mqtt lol
only one node shows up 🤣
Do you hand out contributor badges?
Or maybe potato stamps?
Challenge chips?
🤣
accidental bug find xd
Few low level enhancements like alerting for noisy telemetry nodes and access to packet debugging
`
[messagingSettings]
Noisy Node Telemetry Logging and packet threshold
noisyNodeLogging = False
noisyTelemetryLimit = 5
Enable detailed packet logging
debugMetadata = False
DEBUGpacket = False`
These are currently only sysop log no RF alerts
Uses the config.ini to set favorites on the bbs node, automating the lock of important use nodes in the db
Eliminates a step or two for needing remote admin
(This is just a fancy use of the CLI set)
So... Sorry for asking something so basic, I just discovered this Bot/project and would really like to try deploying one for testing. Am I correct in assuming that I can connect a Pi to a mesh node using a USB cable and have the "Bot" listen and communicate using the serial interface? Or does this require that a Lora hat be installed on the the Pi?
You got it @lunar fog attaching via USB will work.
It’s configured plug and play with 50% of them and the other 50% is ttyUSB0 typically
Thanks for replying so quickly.... Pretty excited about your project.
Yea no problem, hope it’s fun! @lunar fog
It will be a learning experience for me and thats always fun. Thanks @gentle kestrel !
So... I hate to be such a noob with this... But Is there a certain OS that is recommended to install on the Pi (Bullseye, Bookwork)? Or does it need to be Ubuntu server? Sorry to such basic question...
Pretty certain this should work on bookworm (I think this is current stable, bullseye might EOL). No need for Ubuntu server (although there’s probably very little difference between these two)
Thank you!
@gentle kestrel , you are a rockstar... Your utility Bot, is about the coolest thing I have seen on the Mesh. I need to buy a "spudgunman" t-Shirt...
Haha thanks @somber condor
Let me know when you get your Etsy store up.... I'm first in line 🙂
oh spud, THAT last one is ... dangerous as hell 🙂
Bah ha ha .. I'm going to go hunt down your node now... {{evil laugh}}
Just so I can go fix this in your github 🙂
You ever go by PDX Hackerspace?
I think Meshing-n-coffee are in order.
If you see @modern agate (I’m assume they must be in PDX) in your travels buy them a croissant or coffee on my behalf. Between PDXLocs and Spud I’ve learnt a metric ton about coding stuff in Meshtastic . I’ll pay you in hops or smth
You can fill the working directory with batch scripts or whatever to do whatever including remote uptates; this is disabled by default
It’s only via DM so there is ok security all considered if you choose to enable it
If you do something like run nano, it will be a mess so, don’t.
python -m contact
Haha yes, gonna be a mess
not only will I mangle the i/o, i'll break the bot's link 🤣
Well it should be understood this is as said is dangerous as hell
remote meshtasticd update script could be handy tho...
Yea handy for that and other basic troubleshooting; but ideally in scripts with good error handling as it’s not interactive console just one time data burst
Need an api or something back into the bot ❤️
What for? I have some ideas on this but haven’t moved on them
Like I run a script that does something that will take some time, afterwards it outputs "Success on job xyz!" to /tmp/meshtastic.nodeid and the meshbot sees that and sends the contents as a dm to that nodeid, then removes the file
Or just some way for an external script run from anywhere, like a cron job, to put a message in queue to be sent out by meshbot
Just spitballin 🙂
Turn wifi on and off
I like it! To extend upon that, how about turn cellular modem on update things, turn cellular modem off 🙂
So am I there is now a New tool
https://github.com/SpudGunMan/meshing-around/commit/2d44faac98bb977d9b3b9f840ca4dfbec71af3dc
Needs a [bbs]
# enable API script access (increases disk i/o) bbsAPI_enabled = True
Awesome! Gonna tinker with this later
You can easily do that with the new x: command
Have a wifiup.sh +x
And run x: wifiup
Could also x: rm / -rf so tread lightly
Could also x: python3 script/injectDM.py -s 4258675309 —d 1234567890 -m “hello world”
Lotsa’ updates, around bbs messages and telemetry
Better bbslink and detection sensor in sentinel
The bbs API enabled allows ssh of bbs database and it should synch bbs better now
Looks good!
I've got a spare node here ( Heltec V3 ) and a Pi, I'm doing CI/CD builds from your source repo and playing with it 🙂
I've started a REALLY big project on my part -- implementing a TinyMUSH ( already have tiny nethack ) -- but I'm following my fellow PDX'er as you make meshing-around 🙂
Hey, since I was already doing it for the meshbbs code, I went through meshing-around looking for memory/stability issues and improvements. I put in a PR
Woah thanks will check that out
PPSt -- added a second one this morning. Silly, but fun.
Yea rad thanks will check that out!
tips hat you're welcome.
I'm impressed by Copilot's recommendations on this PR
Rather I should say it's a solid-looking PR and I'm impressed by its recommendations
snicker I sense the sarcasm, @mortal shale 🙂
I'm OK at writing technical documentation, but I still lean on Claude Code to clean up things. I grew used to "pair" programming through the '90's and early 2000's.. and twinning with Claude really does feel like working with a team again.
You are very welcome. I thought it would be a silly, fun thing to add
I didn't make the AI perfect... what fun would that have been?
State engines are fun
I remember when spud was first getting started adding stuff, I think there was a joke in there somewhere about the only winning move...
In all the cleaning the messages function went bad. Bug fixed 🪲
My you’ve been busy 🤙
can’t wait to play with this new stuff
survery, readnews, and quiz start -- how are those being activated???
Enable in config.ini games
Readnews is a filemon
Enable enable
Pi 4- update choked...
# git pull Updating ce317d8..3ce24fb error: Your local changes to the following files would be overwritten by merge: .gitignore README.md config.template Please commit your changes or stash them before you merge. Aborting
Haha you might have a older copy it now should say - Git pull resulted in conflicts
Would you like to ..
git fetch --all git reset --hard origin/main
Of note; I did add a speed bump for news.txt
It got moved to data/
I could use more ideas for leaderboard. I have a few like “most messages” and such
If anyone thinks of a funny one
I had one idea for most words used on mesh, vs just most meshages
A GitHub user had most common word
That PR for language file adaptation likely will be some day, that specific PR is not ideal
I will ponder, it would need a lot of extra and be only as good as anyone who uses same ID on them all
I do-- wh6gxz is on many of my nodes-- as many hams have -- just a thought 
how do access quiz or survery?
If enabled should be in CMD?
They each show up but only work via DM
Like a game
(Needs a captive user)
I haven’t finished a admin detailed yet, possibly - ya know documentation..
If the bbs admin says go it will q: start
I think quiz ?
Also has it all
Survey is just a basic game like ham quiz
Quizmaster, however won’t stop playing till reboot or told
.. should add an auto start for it
But; the idea being if you were to use this for a quiz night
I got more ideas for it just wanted the basic ground work in
Survey thing was request from here/ny
fun idea
I also deleted my dev pi folder by accident. It had all my dumb game ideas - I may have thought of a way to get uno but need to find its backup.
I joked to Martin about battleship and had a bad copy because it’s hard to fit the fun in a window of 5x5 or so
I have thought about a text only could be funny
This is very cool dude-- meshing-around sure has grown over the short time I've been using it!
B4
Hit, I think, yes you hit a sub
Glad it’s fun, I’m enjoying it!
Thanks to all the community 🥔
Thought you’d like this Spud https://mastodon.social/@TomasKuchta/115345605766994924
I have finished the battleship implementation
I''m making a port for meshing-around, so that both meshbbs and meshing-around will have the same game/interface if you like
( the idea being inter-bbs play someday )
I increased the grid size slightly.
This was the original design:
Enemy Grid:
A B C D E F
1 ? ? ? ? ? ?
2 ? H ? ? ? ?
3 ? ? ? ? ? ?
4 ? ? ? H ? ?
The problem was keeping the grid size < 200 chars
because we don't have guaranteed message delivery
Teh #$% of using unicode is that each one eats up TWO BYTES .. grrr
So using unicode characters for marking ships, though cool looking, costs
ASCII does have some nice line drawing chars though
BTW, if you have an idea how to include some command keying or persistence of messaging that could be used by the apps for e.g. menu item selection, or even simple stuff like the above battleships board, shout up.
What's holding it back is being able to agree some kind of a spec for it.
I suggested using another invisible character like the bel, e.g. the data link escape character, maybe, to mark command characters.
It's a tough nut.
persistence of messaging was easy, already implemented across meshbbs. There's a backing store, and you can read/write messages that way.
I mean on the client apps - maintaining some elements of the screen, like partial refresh.
That's already implemented in Meshtastic
but in our own clients ( I don't have one, I rely on Meshtastic right now ) that's always a good idea.
If we come up with a common "game and fun" client for meshtastic, we'll have to independently keep up with the protobuf spec and protocol changes. I'm up for it, but it's a HUGE lift to maintain a client like that across iOS, Android, Linux, Windows and MacOS
At that point you're almost better off writing an Electrum-style app ( like Slack or Discord ) that's basically a frameworked webapp
Again, I'm far from /against/ that ... but meshbbs is currently tied to the "tiny, light interface" philosophy
So… I’d like to throw out the idea of being able to pull down an RSS feed (or multiple) for news…. It would be great if there was an automated way to get daily headlines and maybe allow users to click on them to open a link to the story details at the site that published the story…. Just a thought.
Hmmm,
It’s hot off the press
You wouldn’t have seen it till now
You mentioned the link; I have it just didn’t put it yet lmk if this idea is cool or needs refining
URLs don’t travel well always
I did some play testing and the issue was for me, it was hard to keep track of a large game board and I didn’t like 5x5
Ahhh… Ok…. Very cool
Like let’s say you hit f5 I want to see 5x5 around f5
But I can’t remember if I also hit on f19 because I can’t see it from f5
I would need a paper and pencil to keep track(which is fun) but I want it 100% in mesh app (mobile:tdek)
Also; one thing I learned from uno and it derailed me.
How do you get that much data into the mesh well
RSS feeds are getting very rare these days ( RSS is vastly deprecated, many sites don't support it ) but there are still many sites that support feeds.
As long as it's over DM, I don't see why not.
I'll throw it onto my TODO list for meshbbs
@gentle kestrel : In the best of all possible worlds, what would a "tiny battleship" game field look like for you?
Also, I may have mentioned this before -- but since you're in PDX, we should meet up for 💻 🍺🍕
Agree but it’s basic xml which could be handy for many things and if the stack is loaded..
Furthermore @somber condor what I do is have a LLM condense into news.txt which really helps to offload the whole news stack 📰
I’m in NASWI, could head that way some day
Same here, but the number of API calls to either OpenAI or Anthropic can add up $$ if people really pound on it.
Why mesh-bot uses Ollama!
Right now Google is making Gemini 2.5 access free for up to 1000 api calls/day but that could change at any time
I’m trying to long term make the local LLM tool call and use the local wiki for truth table
Ollama + 2.5b models do run on a Pi5 -- passably -- well, but they are all very minified.
laptop isn't bad though, and of course a dedicated miniPC or desktop is ideal
Meshing-around .. not likely gonna save the planet but it may pass the time in a fun way
same for meshbbs... not looking to save the 🌎 ... just looking to let people have fun on the mesh rather than just saying "hi" all the time
Ack
@gentle kestrel : If it help, I do have a Reuters API key ( long story ) if you want to be able to grab well-curated journalism articles from subject queries, and top X news headlines/articles from the news.
So… just as a comment. We are finding in our location that there is an explosion of nodes coming online. But once they appear… no one is using them
Happy to share, it's good for ~50k API calls per day
Right, there's nothing TO DO... which is why meshbbs sends an 'ident' call every 30 minutes announcing it's existence, and a hint that there's more to do.
Ideally it’s a 1:1 remake of the OG but; the catch is only showing where you hit (make sense?) (at the same time showing enough play stats) I know your at a lower level but even with ack on packets needing a retry costs multiplayer
I really like Meshing-around for its ability to push data to end users (weather, alerts, news, etc). It turns the mesh into an alternative information source…
Thanks 🙂
Games are cool too…😜
Nice! I’m just webscraping today I don’t release it because news - politics and I’m full of it most don’t enjoy hihi
I figured with vibe coding most people would enjoy the challenge
I see the news.txt and alert.txt as primary api for the bot
I’m gonna hit it,.. night guys…
I’m very curious to hear about someone using the quizmaster function at scale
I would like to add a html page for it if people think that would be cool
Depends if it works or that cart should get behind the horse
It’s on the list of things that might be cool
Maybe I'm not making a whole lot of sense. 2 elements to this:
- have the app latch on to "hooks" in the markup or whatever in the message (like hyperlink but hopefully lighter) and then present those like the quick-chat options. Mock-up below.
The other element would be to solve the issue of not wanting to resend a huge grid (as given by Spud)
Multipart messages are probably doable within the existing apps.
Partial refresh sounds fun...
It would be cool if we could access this via a message sent somehow to present suggested auto complete words to the mobile client
Client needs to cache and ack part well
5 parts delivered part 3 missing please resend
Once we have -truly- reliable messaging in the Meshtastic protocol... it's coming, but slowly. There was a good proposal in Routing 3.0
I hear ya, this was mostly just for fun
Don’t forget licensed mode
I discount that mode .. most ( almost all ) meshtastic users can't access it.
They just flip a switch
It only drops encryption
And a lot of extra bits
You have maxpayload power
…. Don’t tell me you are still crawling at lvl1, I would hope by now you have unearthed a waypoint and it had a password to the mini game connect4
RSSReader enhancement.
Multiple feeds
Multiple feeds
Possible
Possible?
the issue with the lvl1> prompt is that the main BBS code is always overlaying the prompt even in the games. sigh Really not happy that early decision of mine is biting me in the ass
I'm thinking of just dropping the prompt altogether and saving the bytes
Add favorite helper script enhanced for new client_base automation of favorites
Basically just a fancy CLI for automating it
been playing with readerrss -- here's some of the output for today...
if I type say, readrss meshtastic I get the same output-- the wiki: cmd is cool. type wiki: meshtastic and you get--
SWEET- run ollama-- then ur @ another level --

Yes the rss reader I am personally parsing and I am a potato. I didn’t use a well constructed import at the moment (femtofox includes) so was waiting to see how much use it might get
Like a feed of redit meshtastic might be icky with my basic html strip
A feed of older websites is fine
As mentioned by others rss is old and abused well
(All this said I have no issues added a well formed import)
If it’s a wanted feature, we can push a change to femtofox or people can just add it
.
Also just fixed a bunch of tabs and bugs on leaderboard I didn’t use an ide and it was looking like crap
So it’s working, now haha
I just had a fly over by some meshtastic people and it was totally chaos on the leader traffic for a bit during the relay
Fixed save on the leaderboard
I like playing with it--- I'll try more variables
the leaderboard is very cool bro- play with it a few time a daZe... hihi
I just updated my femtofox bot-- runs on a channel called HawaiiNet that runs through out the Hawaiian Islands.
...and i run one on a pi4 -- a private channel -- runnung Meshtastic, RNS "Reticulum", and meshing-around... Oh YEAAH With aloha 🥔 man--
just typing 'leaderboard' outputs header 'leaderboard' more detail produced this-
cool stuff
--got work to do-- 73 and aloha
BTW I did get output after a the bot was up and running for a bit- by just typing 'leaderboard" a little diff- output it appears each time. shaka
As data is so will it be.
The only leaderboard command is leaderboard reset by admin to clear runtime
I just had a juicy test work well..
I just noticed that some of the games (dopewars, videopoker) aren't handling game commands. Instead the bot is responding with either a joke or the default response. Could be just me.
Like a new thing? (Could be a memory cleaning problem)
It’s the same game engine for tick tack toe but was modified at the same time, ttt works as will quiz last tested
It can get confusing, I try and keep debug full
Should say “LLM disabled” game
TTT is working for me, so is Blackjack
and hangman?
Yup hangman as well is working.
I will check the game timer for the older games are fresh I would bet they got lost in the shuffle by the sounds of it
But if I choose 'e' in hangman it ends the game
I'm starting to get the shakes without my dopewars
Can you tell me more? I can’t replicate yet
I did futz up the new vox/radio module now fixed
It’s not documented yet
But should .. turnkey
Make sure to git pull latest, also just extended the timer for it to 3 days or so
Will do. Thanks
Ok I found a issue
It's me right, I'm the issue
@gentle kestrel : What is your PDX Node running meshing-around?
I'm adding some serivce-reachability tests to my grafana monitor...
What’s the node ID?
No it was me, very much.
Just got a big patch in to fix
lmk!
Thanks btw Martin for that PR
I should have tested more but the better error handling is welcome
I had a lot of trash
Thanks a bunch! That fixed it.
Thank you for calling it attention
So, just double-checking… should we still be using the same ”update.sh” she’ll script to upgrade to the latest version ?
Yea should be fine
The tic-tac-toe changes are where games went wacky - now back on track. My fault all along.
It's possible that the 'Goodbye' I caused here was from me trying multiple games at once. It's working since your fix.
Lemonstand has always given me this output, so I just leave it disabled. Cuz you have cooler things to work on
Hope it’s fixed now
I'm still getting the same network stress-test output after updating.
Also noticed that the sysinfo command is only returning 'Gathering Telemetry try again later⏳'
Wait longer
Yes you did - I have reworked it some more well a lot
I hope .. all games work now
Yes it is now thanks for idea
Took a little bit to iron that one out
I also added a new fun one
`
voxDescription = VOX
useLocalVoxModel = False
voxLanguage = en-us
voxInputDevice = default
voxOnTrapList = True
voxTrapList = chirpy
`
“Chirpy hello world” 
It works like Siri but it’s chirpy!
Sends whatever it hears to mesh
I’m not sure how stable it is; but I let it go 30min and it still worked
Yelling hey chirpy time for dinner
🎙️Detected VOX: it's time for dinner
I may have found a bug (or maybe I did something wrong). To put this in context, I send out the days weather to our local mesh every morning at 7am. I use the advanced scheduling functionality. In my "config.ini" I have a location defined (which is our local Intl Airport), but in the node that I am using for "MeshBot", I do not have a location defined (i wanted to keep the node off the map). For the past couple of days, I have been getting this error when the mesh notification is attempted to be sent ("2025-10-12 07:01:00,880 | WARNING | System: Location for NodeID 0 not found in nodeDb", "2025-10-12 07:01:12,250 | WARNING | Location:Error fetching weather forecast from NOAA", 2025-10-12 07:01:12,250 | INFO | Device:1 Channel:0 SendingChannel: error fetching data").
This morning I set a fixed location in the node and the scripted mesh notification went out correctly. So it appears to me that the advanced scheduling functionality is not using the location set in the "config.ini"
Just a headsup...
Oh interesting, yea thanks
You had too wait too long sorry haha I was .. yea fixed now
nod I've been having issues with getting weather from openweather, so I ended up implementing a 10 minute "cache" and if there's a miss after 10 minutes I just say "Weather not availble" so the data won't get stale.
NOAA has some rate limits in place that seem to soft-block for a while if you hit them too hard
Yes they do
Thanks for this. Dropped a new release and a significant change to the location behavior to include fuzzing
Corrupt data or just rushing to new code. I also find clearing nodeDB helps things sometimes.
How patient do you expect a dopewars junkie to be? 😂
Thanks for the many fixes! You da 🥔
If you use sentinel I would be interested to know thoughts
Thanks!
I feel it’s important to document this milestone
PPSST - fixed the danged "lvl(X)" problem in MeshBBS 🙂
Even have the MUSH working more or less reliably 🙂
So, this morning, none of my scheduled events fired. The scheduler is enabled and the events are setup. I don’t see any errors in the log file. I can get the data (weather) via DM, so I know that meshbot is working.
Is it from like 6+ hours build?
Yes sir…. Updated yesterday evening..
I found a issue at midnight
Is there a version baked into the code somewhere?
Not yet I should do that I hope I have ironed out most of the blocking code
The addition of an idea for sentinel needed handled better
So if you happened to land on that it would block the watchdog
So I had enabled the high altitude feature… I must have received 50 notifications overnight. That continued to work and I had to disable it this morning
for over 2k above sea?
Or you saw logs
Anyone can obviously fake a packet and such but - the addition of the flight tracker should help narrow it (I get one or so a week and they match up for alt typically)
So, I’m in El Paso Texas… we are at 3500 above sea level. But yes, it was set for 2000. We have a military base here and I think someone had a Seeed 1000 in there pocket. I kept getting the same notification over and over. Like they were in a flight pattern around the city
I can send you logs if you want?
Yea, also you can bump the ceiling
Altitude in meters to trigger the alert
highFlyingAlertAltitude = 2000
I am not skilled in actual flying so I have no idea I just thought 2000 was a nice default if anyone has better it will be changed
I find this feature to be one of the most fun
I am also on the setup for approach for Seattle some days and I see a lot of flying hellos
Also Sysinfo was fixed in the fix of the watchdog as well
Sorry, got pulled away.... So. Should I do another update this morning to fix the watcher issue?
So.. I'm not sure if you are interested in this or not. But I will send you a copy of the log entrys for "High Altitude" alerts that started yesterday for my system.
If you look at the last 7 entrys, they are all the same, just different time stamps
they were all sent out to the mesh...
Ahh yes, sometimes for any reason someone could have a bad sensor or bad data
list of nodes numbers to ignore high flying alert ex: 2813308004,4258675309
highFlyingIgnoreList =
Can also block them outright but
This value assumes just bad sensor
I have 2 at least near me
Same way stuck at some wacky value
Also I could use input - I was considering narrowing the flight return to like 300m of the node alt
Wasn’t sure if that was handy or not as fun
so, a question. all of these events are being report by NodeID:382705832. This is the main "router" that sits atop a mountain that El Paso wraps around. Unless I am mistaken, we are not identifying the node that is actually on the aircraft that is (I assume) pinging and being discovered by the local node.
Is there a way to identify the node on the plane that is being "discovered" or reporting its position via telemetry?
Also, I just pulled down your latest code.... I am getting this error when I restart the service : "2025-10-14 08:33:47,130 | WARNING | System: No schedule.Value set edit the .py file to do more. See examples in the code."
The scheduler service is enabled.
Also the watchdog does not appear to be running...
or responding
Did you enable the request Position feature
Disable it for now
Bad location to log this fixed
It was just a log line
This is a bug @somber condor
can you pull a new one
Major refactor to scheduler
Put it in a module
So if anyone using a custom 🚨
Pull a new and have some peace of easier edit
Pulling now
Ok, just completed the update. I really like that you moved the schedule stuff out of mesh_bot.by and put it into a "modules" sub
I am rebooting the machine and I have a scheduled event in 4 min
I will let you know how it goes
so... mixed results.... The event that was scheduled at 10am my time fired correctly (success). I had a second event scheduled at 10:01am "weather" and I see this error being logged. "2025-10-14 10:01:00,424 | ERROR | System: Scheduler Error name 'handle_wxc' is not defined"
Maybe a typo?
No this is why I hadn’t moved it all yet le sigh .. it might need to go back
If I just run the wx command... it works
I looked at your code... I can;t see an issue
I need to go out an run some errands,,, I will be back later
Roger - I think I got it working. Thanks again for calling it out
@gentle kestrel : how effective is a ban-node in a world where the firmware can be patched to change nodeID?
About as much as being in a open network. I use it to quiet things mostly
might auto block some day depending
.
“Hey chirpy” feature seems more stable now
This demand on resources sent me all over thanks for patience to anyone disrupted
I just had a memory leak issue ( first one! ) on MeshBBS, so I feel your pain
I couldn't get a full 24-hour test run in with the TinyMUSH code loaded in
Would not have noticed it if I hadn't been running on a Pi2ZeroW
A normal Pi4/Pi5 would have taken 3-7 days to fall over
Actually, since I have perfmon and memory guard stuff running here, is meshing-around in a state you'd appreciate a 72 hour run to get stats?
the last time, I just went through the code looking for easy "gotchas" but this is a bit more instrumented
I'm doing a valgrind run on current code.
You can now say
“Chirpy tell me a joke”
I’m thinking about throwing my hat into the bbs game. One caveat I need to get gud at Ruby (for work) so will try writing it in that
you can .... there are a lot of Meshtastic BBS/bot frameworks now though, so unless you really REALLY like the pain...
I'd extend the ones that already exist, like meshing-around 🙂
mine ( MeshBBS ) is itself a port of a much older codebase from the '90s
Which started life as Pascal, then was ported to Java, and now Rust
Also a matching (working example) https://github.com/SpudGunMan/meshing-around/blob/main/modules/udp.py
The .ino works but I can’t figure out how to print decoded payload yet
(Likely need a key somewhere)
Let me take a look.
I'm pretty good with arduino code and such
I'm assuing the arduino code is the broken one?
Yea
do you want the raw UDP payload, or the decoded Meshtastic protobuf?
I assume the latter?
Yea
Easy. One sec.
Nice thank you
do you have all the nanopb headers you need from the protobuf repo?
You'll need:
portnums.pb.h
user.pb.h
postition.pb.h
Yes I got them
That should do the trick.
Two important things:
Comment out:
// Always print raw payload first
Serial.print("Raw UDP payload (hex): ");
for (int i = 0; i < len; i++) Serial.printf("%02X ", buffer[i]);
Serial.println();
Serial.print("Raw UDP payload (ASCII): ");
for (int i = 0; i < len; i++) {
char c = buffer[i];
Serial.print(isprint(c) ? c : '.');
}
Serial.println();
If you find it annoying... that will print the raw UDP
and ... packet.which_payload_variant == meshtastic_MeshPacket_decoded_tag is important, it confirms the oneof contains decoded data rather than an encrypted blob
Oh, and since it's meshtastic ( yay ) and you're on Arduino ( um, yay? ) the width of words kind of kicks you in the butt. So, positions are scaled integers.
Divide by 1e7 to get "normal" lat/log ( and don't forget to cast to float )
I'm having my own problem similarly -- I want to use WiFi with the BBS, but haven't started serious work in the that direction yet. Serial connection is good enough for now.
wifi or ethernet
im missing stuff like user and position proto gonna need to fix that still
I can build proto now but still stuck
Can’t find user
I also don’t see a position
In local build or in firmware
the .h files should have built for you. hold on, let's see if I can fix that
you should be able to grab them from: https://github.com/meshtastic/protobufs
Totally and I can build them fresh, I’m failing to see or understand where user.pb came from as it’s never in other projects? For example I can see a telemetry but not position
https://github.com/meshtastic/protobufs/blob/master/meshtastic/telemetry.proto
I assume, the existing call needs to be on existing methods
protoc --nanopb_out=. mesh.proto
protoc --nanopb_out=. user.proto
should generate mesh.pb.h and user.pb.h ( and c )
Mind you, I generally only use Arduino for "light" projects, I've never seriously considered using Arduino + ESP32 to do protobuf stuff, even though it should totally be possilble.
These days the smallest computer I use tends to be a Pi Zero W or 2W
I can build everything but that? I don’t see options for it
oh! Looked through.. I have a much older version of the protobuf defs in my Arduino build directory.
Current:
#include "mesh.pb.h"
#include "pb_decode.h"
#include "pb_encode.h"
sorry about that, they moved 💩 around
This makes sense brb
Progress, I think just encryption to decrypt now?
I don’t have a ham lab handy, need to make a net
I can't wait to try this out! 
I run mdup on a pi- running ubuntu 👀
Nice! There is a mudp python implementation and I’m working on some ideas now with pygame!
The arduino variant isn’t currently decoding encrypted packets (beyond me at the moment)
For me, I need to iron out in Python first anyway
The idea for the .ino is wide use by community for boilerplate ESP32 display for tinkering
The mudp variant will work fine for the bot here
Ideally I would like to have it on Esp32 only and it could be an AP
Connect node it shares UDP and you have a wireless/wifi display
So you can do wild things like make a full game
Or have a fun boilerplate for a custom weather display
I would like to have a bot dash on this as well
All just very rough drafts
All this to say mudp is using Python in modules folder 📁 UDP.py run it Python modules/udp.py it’s just a demo right now
The .ino in etc/ is not fully working it will only pull out clear text
To from junk
I think I know how to decrypt the packets. Let me see if I can implement that a little later this afternoon
Spud my good man, is a UDP-only connection supported now? We (@eager zephyr and/or I) may finally be setting up an Oregon meshing-around soon.
Not for lack of nodes....
Too many bouts of dysentry?
I know nothing about Oregon, except what organ trail taught me
It should be, perhaps soon!
Currently the bot rides on firmware(using the api to access the nodeDB on device)
while a UDP will need a database and all, firefly base looks nice 😉
The advantages of each I haven’t landed on which is better, Linux native can hold a lot and keep up with changes
ultimately holding rev1 on existing concept or flipping to UDP for v2 with a new core
The PR to return nodedb as json may force this as well
It's accurate.
We're running linux nodes but don't want to consume the API so we can still use clients for messaging. I was thinking we could also run a 2nd meshtasticd (why recreate the wheel) and mesharound with that, while the daemons have a UDP party.
Indeed, this is why it’s been not quick to release, having it in a vm -meshD instance you can remote manage the node /infrastructure and there is no need for the bot to have any keys or data of consequential value
The negative is an added hop but so is UDP
We all run custom firmware with hops set to 11.
My initial thought is, I have message ID from firmware and if I see it on UDP ignore it as a dupe, I need to decode the channel ID still from UDP
Allowing disable of a classic interface and the bot rides on UDP alone as a “default install”
I have been thinking about using senddata() to back channel graphics for games lately via UDP
ESP runs a display manger
Bot sends command, control via data line
So let’s say tick-tac-toe was running text like it is now
Add a UDP display and you have rendered whatever via the “display manager”
simple as Senddata(“battleship:d4”)
The text_message sends in queue with _data
Giving the illusion of a unified device
While that idea bakes, another I couldn’t help but wonder if MartinsMuSH can expand into massive multiplayer Oregon trails. Or sim city 1.0 clones
Like if there was an esp with a mouse or touch screen
It would be a tablet for mesh kinda
Advertise as an AP
if I drive past I5 PDX and my node appears on mesh do I become a new NPC if I say hello and have a tablet on me do I collect a Pokémon?
Further ideas, what if it’s a $30 device and I have 50 of them to hand out in a emergency to full out a survey
sendmail(me@winlink)
Emergency.... Survey?
Whatever, what if it’s an exit survey at a family reunion or a picnic for the company
It’s common in USA emergency to have surveys of data
Winlink is a off grid email system with such a system of forums and data
Used by many in FEMA aka USA EMCOM
Survey idea came out of a snow ⛄️ request
If you enable survey and say survey snow
This was a direct request from user community
(I hope I nailed it)
Need some snow first 😉
Ok, that makes sense
FoxInSocks 🦊 🧦
Other things this architecture allows is stuff like the screen responding and the bot detecting an attached screen and altering from text output to video, giving display parameters
Need to test with pygame zero
Could also add a basic game pad
Offloading the demand (or putting it on mobile) for such an inclusion 🕹️
Shows you any missing config values
Script as part of update.sh or run from script/
(Run from project root) python3 script/configMerge.py
You don’t need to add them if the default is what you want
@gentle kestrel Just had an idea for a potential work around to add UDP support. Maybe this is crazy so feel free to call me dumb. (Or maybe you’ve already thought of this)
Second meshtasicd and the bbs in a docker container. No actual radio for the “virtual” node, UDP only. BBS connects serial to this “virtual” node. Virtual node communicates with primary node / meshtasticd over UDP for transmission to actual mesh.
Theoretically allows any number of these serial based softwares to be packaged and run together on one Linux native node and each have their own virtual node IDs and none tie up the serial connection of the real node.
Also don’t have to reinvent the wheel to add UDP support, it basically relies on the built in UDP of meshtasticd. And you get all the same features of connecting to a dedicated node, in theory.
Yes this is a very effective way to build a network today!
You can also use MQTT this way
I don’t even develop with a radio anymore 100% using mudp and meshtasticD
True, but with MQTT you might lose people depending on configuration on both ends.
Nice, so it is possible!
Yes just that the architecture is possible. I will likely have ability to listen and respond to UDP native soon; it may be a core replacement, still ironing out details.
(I just got my first linux native node running yesterday so I'm a bit new to it)
Sweet! You can enable UDP on it and it will respond via the real radio but be advised any channels with UDP may be uncontrolled data
Right now; the bot has no channel ID for UDP
I lost power likely for the day so had to power down. But I was working out a lot of core issues around random bugs thanks to community 🙂
Anywhoo nothing huge but the update.sh got some enhancements to include backing up your data/ files as well as incorporating a log of any config.ini or features you may not know of
You can git pull and update and then bash update.sh
Or run update.sh twice to enjoy
Did fix one game play issue with core dm
Likely broke survey haha
But .. I didn’t update my offline bot before powering down to deal with the day
I lost all my high scores recently I cried I had some insane coins going
Lightly tested but seems to work for multiple d's using different ports https://github.com/SpudGunMan/meshing-around/pull/219
(when your power comes back)
Nice thanks, it’s in along with a refactor on game tracking (better game management now)
1.8 release, I hope I’m nearing the end of the bug 🪲
1.8.1 with new hop logic and better MQTT and IP nodes
A reminder there isn’t full UDP channel support yet. I am unsure how to get the channel hash out of the current methods used or if I need to bring in new ones?
I’m interested if any negative impact of hop change and specifically which firmware may be negatively impacted
users of UDP or MQTT likely the most
I do dev on latest alphas normally
Released pushed a lot of stuff but this is fun
https://github.com/SpudGunMan/meshing-around/commit/1324f83f1785b6750771e2b6a66804cee20cdec6
busy 🥔 ! i'll 'git pull'

...and one on femtofox --- oh yeah!
not getting motd output... used to-
there is a HUGE update coming
What the heck - fixed now
i'll have to play with the survey stuff-- get a few people to participate.
MOTD is back
You can also run sysinfo as an admin from a DM it will give you local IP guess and public IP guess
(With update)
Sentry got a overhaul
# list of watched nodes numbers sentryWatchList =
Will alert if they are seen outside the zone
Question - is this thing supposed to be telling me jokes every time it doesn't recognize a command? lol
Yes that is default way to fill space and time 🙂
aloha the cmd mwx stopped working hmm.. no error output -- nada -- wx is good and other weather related cmd's 🤔
Doh! Thank you. Fixed.
I did add what i think is the api key, but maybe not? I'm not home now to copy exactly what I entered. It doesn't give an error, just keeps giving an example of how to use it
This is its basic use
Register for free API https://www.n2yo.com/login/
n2yoAPIKey =
Paste key no “
You can also be fancy
Or stack,them,in .ini
I patched for a 2.6.11> issue to hopefully allow direct pings back and not allow as much of a mess with UDP
Update firmware 🙂
OK, we now how meshin around up and running in Portland and in Salem and will soon have them linked up to sychronize. I only have one question so far:
What? https://github.com/SpudGunMan/meshing-around/blob/4aa65dad6acb798829ee4a40adfc3c62b72be4c7/modules/games/joke.py#L58
mahalo spudgunman for meshing-around!
I should add a pylint ignore huh?
I sent the best bots on this one https://github.com/SpudGunMan/meshing-around/pull/223
I was just browsing variable values to actually fix the ip:port support. Got it working for real this time on a sim linux node alongside a "real" linux node/radio with UDP enabled
Thanks! update in with better hop debug logs now 🪵
I'm getting an "error fetching data" when I try to run ea on my new install, any thoughts?
Increase the urlTimeout in general.ini
Increased to 60, still failing (instantly)
Just pushed (so update) a better error log with result code
Will give like 404
(Always dns)
Updated but it doesn't appear to be any more verbose.
I added much verboseness
I'll try it again, hang on
None configured, should be a pretty straightforward connection to the internet.
Let’s hope you got a pool member
Is api under open government budget?
Ping @thedonald
The SSL looks valid and curl is able to connect from the same device. Weird.
check out https://apps.fema.gov/IPAWSOPEN_EAS_SERVICE/rest/feed
thats whats used
is your local time wayyy off
Fixed by adding this to the [service] block
in mesh_bot.service
Was this a raspian or armbian build
Cool thank you. I will add this vs ignore them
Got that in for new installs I hope to test soon
I have been meaning to add this for some time
Excited to hear if handy
Ping and hops might work better, specifically on UDP hosts
This is what i put for the api key
I guess i don't know what an api key is. I thought i signed up for it.
It will look like a windows activation key and you ideally don’t want to share it
On login,
Okay I've been scrolling up and down and I must have missed it buried in all of the ads LOL. We'll check again
Found it
I will take ads’s in html for free api 😋
👑
Hey so we have bbs's linked up in Portand and Salem with the scheduler set to sync. Did you have any plans/thoughts about pushing new messages to other bbs's when something gets posted?
Currently it was a pull request only, if you ssh:copy you can do it live by enabling the “api” which is just file polling mostly
The reason for pull only is it’s administratively ran at midnight?
I can otherwise post a bunch of messages and crash things mid day downgrading high use time
I do have a plan to put compression and back haul on port 256, can send the whole package 📦
The bones are there just not the testing yet
Thanks to community ideas a throttle of “only every 30 min” is now on 🚀 high flying
So same flying node won’t fire without 30min delay
ditto -- I have the beginnings of a "republish"/"sync" setup for meshbbs, but it will come down to handle async replication
I may just "borrow" from the best -- usenet
UUCP worked, for a reason
You mean a pickle file 🥒 and ssh.copy isn’t good?!
It's great! Don't get me wrong
but that's if you have each node trusting the other 🙂
<-- information security, remember?
They share PSK channel
Tru
This means someone hacked your coms and the data of the pickle!
So… my instance works great when folks DM the meshbot. I am still trying to schedule a daily weather forecast. If anyone’s runs the “Wx” command, it works perfectly. If I run it from the scheduler, (using your example by in commenting it), I get “error fetching data”. I can run the manual Wx command 30 seconds after the error is posted and it works fine.