local ai cannot edit

[anubissbe]

When using llm studio and load any local model (qwen, llama, gemma,...) the is on agent mode, but cannot write code to disk or execute in a shell it

[Mr-Philipp]

using a macbook and ollama installed locally i have the same issue. maybe relevant: https://github.com/voideditor/void/issues/419

[cetindogu]

+1 same issue

[beldar]

me too, tried with devstral and couldn't resolve file paths and with qwen3 I kept asking to edit the file but only gave me instructions on how to do it myself like it couldn't access the tools or something

[waylonwang]

+1

[konmeo]

+1

[andrewpareles]

Thanks for reporting; without more details I would guess this is unfortunately a model intelligence problem... small models struggle a bit with calling tools, but the next round of OSS models might change that.

[nablitalksdev]

+1