Using LlamaBarn with a already running llama-server?

[gsgxnet]

Since a long while I am running llama-server on this system, serving gpt-oss 120B using SSL and API-keys. The web frontend is good for using llama-server from other systems within the home LAN. Locally on this system I would like to have a frontend with more features and functions for accessing llama-server. LlamaBarm might be that frontend I am looking for, it seems it can not use an other llama-server instance running independently from itself.

So my feature request would be to make LlamaBarn configurable in such a way that it does not depend on running its own llama-server instance.

Thanks for all the very useful open source software from the ggml cosmos

[erusev]

Thanks for the feature request! I want to make sure I understand what you're looking for.

LlamaBarn is primarily a model manager that handles downloading and running local models. It doesn't have its own chat interface; the web UI you see when a model is running is actually the one built into llama-server (from llama.cpp).

Could you clarify what you'd like LlamaBarn to do for your existing server? For example:

• Are you looking for a native macOS chat UI (which LlamaBarn doesn't currently have)?
• Or were you hoping to use LlamaBarn's catalog to manage models on that external server?

Knowing more about your workflow would help us understand if this fits within LlamaBarn's scope as a model manager.

[erusev]

I never heard back about what you'd want the app to do with your existing server, so I'll close this.

The app runs its own server -- that's most of what it does -- so using one you already have isn't something we plan to support. Happy to look again if you're still interested.