commit
Generate a Conventional Commits v1.0.0 message for the currently staged changes via a local LLM.
How to use it, with examples: AI assistance.
All options
| Option | Value | Default | Description |
|---|---|---|---|
--provider, --backend, -b | TEXT | LLM provider: "managed" (local llama.cpp run by fobis, the default), "ollama" or "openai" (--backend is the former name). | |
--url, -u | TEXT | Base URL of the LLM server. | |
--api-key-env | TEXT | Environment variable holding the API key, sent as a Bearer token (hosted endpoints). | |
--model, -m | TEXT | Model identifier to use. | |
--max-diff | INTEGER | Maximum staged-diff characters sent to the model. | |
--refine-passes | INTEGER | Critique-and-rewrite iterations after the initial draft (0 = single pass). Recommended: 1-3 for small/fast models. | |
--apply | Run git commit with the generated message after review. | ||
--config, -c | TEXT | Path to a custom FoBiS user config file. | |
--show-config | Show effective LLM configuration and exit. | ||
--init-config | Create a commented default config file and exit. | ||
--show-facts | Print the facts of the staged change (scope candidates, Fortran line kinds, type hints) and exit; no model is run. | ||
--no-facts | Do not compute the facts of the change (free scope). | ||
--json | With --show-facts: the facts as JSON. |