Repository Context, How does emote embedding model provider works?

[rudiservo]

I didn't understood how this works honestly, does tabby makes the context, or do we have to make it in the external provider, like llamacpp?

[wsxiaoys]

I am about to give a talk on how context providers work internally, focusing on their usage in code completion and answer engines. I will share a version of the presentation deck with the community shortly after it's ready

[rudiservo]

I searched and I think I understood what it is, the 3 things I don't understand is:
1- if you use external completion does the embedded still works in tabby?
2 - the scheduler command is gone, how do we manually index?
3 - how do I index with llamacpp?