With codebase indexing is it possible to instruct kilo on which different projects are related and which aren't?

[cstroliadavis]

I'm still really new to this, so this might be a stupid question. Please forgive me.

Most of my current knowledge is based on what I think I understand from the docs and not actual experience.

It appears as though codebase allows searching across projects.

I envision using this for lots of different projects I work on as well as POCs I'm writing and more. I'm curious if this will cause issues if I end up with lots of unrelated projects or bad ideas that never get implemented being considered as factual documentation when used.

Is there a way to ensure that when I'm using this on various projects that it only considers certain other project indexes to be related and ignores others?

• Would I just use a different qdrant key or something? would I need to set up different qdrant dbs
  • I'm also very new to qdrant, so if there is a way to partition indexes within a single qdrant db and perhaps reference that using the qdrant url like http://localhost:6333/xyz, that would be preferable

[onestardao]

this looks like ProblemMap No.4 (cross-index contamination / cross-project drift).
you can fix it with a tiny semantic-firewall step that scopes codebase queries to only the project indexes you want — no infra changes.
want me to paste the snippet here or share the link?