ollama-commit-mate
is a CLI tool designed to generate descriptive commit messages using AI models from Ollama. It allows for flexible configuration to meet various needs when creating commits.
Note: Ensure that you have Ollama and models installed, as ollama-commit-mate relies on it to generate commit messages. You can find installation instructions for Ollama in their official documentation.
To install ollama-commit-mate
globally, use the following command:
npm install -g ollama-commit-mate
Option | Alias | Type | Description | Default |
---|---|---|---|---|
--model |
-m |
string |
The AI model to use (chat models in Ollama). | llama3.1 |
--baseUrl |
--url |
string |
The base URL of Ollama port. | http://localhost:11434 |
--exclude |
--ex |
array |
Files or directories to exclude from the diff. | ['node_modules', 'dist', 'package-lock.json', 'yarn.lock', 'pnpm-lock.yaml'] |
--untracked |
--un |
boolean |
When true, this feature allows the inclusion of untracked files in the output, providing a more comprehensive view of system changes. | false |
# Generate a commit message using the default settings:
ocm
# Use a specific AI model for generating commit messages:
ocm --model mistral
# Specify base URL for your Ollama port:
ocm --baseUrl http://localhost:11434
# Exclude additional files or directories from the diff:
ocm --exclude 'deploy/index.ts' 'dist'
# Include untracked files in the output:
ocm --untracked