Daniel Nowak

This is great. Any specific reason IQ3_M isn't included?

If you comment out line 463, i.e. `.encode('utf8')` then it works as expected. This patch should be merged in. Actually, prefer the approach `r.decode('utf8')` when passing to StringIO

`We were also downloading a configuration from the server, and setting that as the initial configuration.` Is config.toml still needed to get Local LLM running? I think it behaved this...

Think the concern is the opposite situation If you've already selected a model manually in the UI, will the UI ignore subsequent environment variable changes sent with -e? If we...

Got it, so the current recommendation is to omit LLM_MODEL and AGENT from the environment variables. Those should be set directly from the UI. However, environemnt variables are still necessary...

Absolutely. I think Cypress.io could be handy here. Since there's a backed & a frontend to run separately on the CI pipeline host, [this would probably be a relevant Github...

This change also helps for someone using Ubuntu 22.04 within WSL. In that environment, python 3.10 is the system default and you have to manually install python3.11 If we merge...

Is there already a way to expose Agent-specific parameters to the UI, or would I need to perform a build to adjust this parameter below (which I suspect may be...

+1 for dev container. Any update here? Should reduce effort wasted trying to accommodate different developer environments. Examples of environment-specific issues that (without a unified devcontainer) limit dev toolset selection...

Yes I was thinking about the potential to leak, it's a valid concern. However the approach here is to require an environment variable to be purposefully enabled in to activate...