Skip to content

Add 'ghost --local' one-shot offline local model#19

Merged
0xadvait merged 1 commit into
mainfrom
local-flag
Jun 23, 2026
Merged

Add 'ghost --local' one-shot offline local model#19
0xadvait merged 1 commit into
mainfrom
local-flag

Conversation

@0xadvait

Copy link
Copy Markdown
Collaborator

Local models are opt-in (GHOST_LOCAL=1), but the only way to use one was the interactive /model picker. Add a --local flag that routes a single run to the local model (the ollama-local provider's default_model -- uncensored-local 32B if present, else ghost-tool 7B) via the engine's -m/--provider override, fully bypassing the scrubber + og-veil (true offline incognito). Falls back to a clear message if no local model is installed.

Verified live: ghost --local -z ... returns from the local model and scrubber.log is unchanged across the run (nothing left the box).

Local models are opt-in (GHOST_LOCAL=1), but there was no one-shot way to use one --
only the interactive /model picker. Add a --local flag that routes this run to the
ollama-local provider's model (uncensored-local 32B if present, else ghost-tool 7B)
via the engine's -m/--provider override, fully bypassing the scrubber + og-veil
(true offline incognito). If no local model is installed it says how to add one.
Verified: scrubber.log unchanged across a --local run (nothing left the box).
@0xadvait
0xadvait merged commit 385dcd6 into main Jun 23, 2026
3 checks passed
@0xadvait
0xadvait deleted the local-flag branch June 23, 2026 21:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant