Code execution

[jarodium]

Hello, i'm giving my first steps in AI generated content and I would like to ask for some help.

I want Jan/Cortex to invoke my code execution api and use the results from it in it's response.

Is it possible? Installing open-interpreter and having another software that is running the same model in parallel with Jan seems redudant, so I am looking for a Jan/Cortex solution.

From what I have learned, I need to configure "OpenAI" setting to use an external tool ( for example my execution api ) but, from what I understand, Jan will use it to call also the model right? I mean, I want to decouple Jan local server from my execution api so I can maintain it without the need to code a AI proxy myself.

[Kreijstal]

these people don't care about code interpreter

Librechat you have to pay <- WHAT
open-webui it does work (provided you install plugins), but they live in ancient python 3.12 instead of python 3.13
aichat does also work... but you have to configure it.
People don't understand code interpeter is one of the most useful tools, it should be enabled by default.

[anonymousmaharaj]

ancient python 3.12

lol