Need help to set up a simple chatbot with means of bricks

To get familiar with the concept of bricks I just wanted to set up a simple application, that runs a chatbot with a specified LLM.

I found the example "chat with a local LLm using a specific model". Copied it. But the run button is deactivated. The model " qwen 3-4B Instruct" is downloaded and associated to the program. Already in the python text.

After quit and reopen app lab, the run button is active, Clicking on it, the app reports to be launched and running. But nowhere any window or prompt to enter requests or seeing answers.

Need help, because could nowhere find a solution.

Maybe write your own implementation of Eliza? It was quite popular on 8bit computers with restricted sources, so it may fit in modern microcontrolers :slight_smile:

Does anything "qwen" related show with ps -a | tail (last activities of ps)?

No. Only

"arduino@ventunoq:~$ ps -a | tail
PID TTY TIME CMD
379035 tty2 00:00:00 gnome-session-b
457861 pts/0 00:01:32 geniex
482950 pts/1 00:00:00 ps
482951 pts/1 00:00:00 tail
arduino@ventunoq:~$ ",

as I am currently downloading this model via geniex. Although the app lab shows it as "installed". But geniex does not find it.

I think, I understood the idea of examples, bricks and apps. The intention is clearly an arduino-like IDE. But something prevents it to work as intended.

Update:
After countless approaches without success, I decided to „factory reset“ the device. After reinstalling the ubuntu Linux from the arduino site, the first success:

A, chatbot from the „inspirations“ -list in the app lab, worked as expected. May be the ubuntu updates, that I accepted on the first run, and refused after the reset, have been the cause of the misbehaviour. But I am not shure.

At least the first step is done. Some excercise will be required, to handle the new development environment in the same intuitive manner, as the former Arduino IDE.

Thanks for taking the time to post an update @albrecht1955. I'm glad it is working now.

If you encounter the problem again, just let us know and we'll investigate further.