[#maintenance] v0.2.0: OpenAI-compatible API, local-LLM accuracy fixes and CLI upgrade

.github/workflows/tests.yml

@@ -0,0 +1,26 @@
+name: Tests
+
+on:
+  push:
+    branches: [master, main, maintenance]
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        ruby: ['3.1', '3.2', '3.3', '3.4']
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ruby/setup-ruby@v1
+        with:
+          ruby-version: ${{ matrix.ruby }}
+          bundler-cache: true
+      - name: Run tests
+        run: bundle exec rake test
+      - name: Scan dependencies for vulnerabilities
+        run: |
+          gem install bundler-audit --no-document
+          bundler-audit check

.gitignore

@@ -45,7 +45,7 @@ build-iPhoneSimulator/
 
 # for a library or gem, you might want to ignore these files since the code is
 # intended to run in multiple environments; otherwise, check them in:
-# Gemfile.lock
+Gemfile.lock
 # .ruby-version
 # .ruby-gemset
 

Gemfile

@@ -0,0 +1,3 @@
+source 'https://rubygems.org'
+
+gemspec

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Afonso Neto
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,6 +1,8 @@
 # AI git commit message
 
-A Ruby gem that automatically generates concise and meaningful git commit messages using AI models via [Ollama](https://github.com/ollama/ollama) or any OpenAI-compatible API.
+A Ruby gem that automatically generates concise and meaningful git commit messages using a local LLM via [Ollama](https://github.com/ollama/ollama), LM Studio, llama.cpp `llama-server`, or any other OpenAI-compatible API.
+
+![Terminal demo](docs/demo.gif)
 
 ## Installation
 
@@ -26,13 +28,13 @@ bundle install
 
 - Ruby 3.0 or newer
 - Git repository
-- [Ollama](https://github.com/ollama/ollama) model running locally or other OpenAI-compatible API endpoint
+- A local LLM server: [Ollama](https://github.com/ollama/ollama), [LM Studio](https://lmstudio.ai/), [llama.cpp `llama-server`](https://github.com/ggml-org/llama.cpp), or any other OpenAI-compatible endpoint
 
 ## Usage
 
 ### Basic Usage
 
-Generate a commit message for your staged changes:
+Generate a commit message for your staged changes and commit with it after a confirmation:
 
 ```bash
 ai-commit-message commit
@@ -43,18 +45,39 @@ This will:
 2. Review your recent commit history for style consistency
 3. Consider your current branch name
 4. Generate an appropriate commit message
+5. Ask for confirmation and create the commit (answering "no" keeps the printed message so you can edit or copy it)
+
+### Message only
+
+If you just want the suggested message without the confirmation prompt — e.g. to pipe it somewhere or review it yourself — use `--message-only`:
+
+```bash
+ai-commit-message commit --message-only
+```
+
+### Conventional Commits
+
+Ask the model to follow the [Conventional Commits](https://www.conventionalcommits.org/) format (`feat:`, `fix:`, `chore:`, ...):
+
+```bash
+ai-commit-message commit --conventional
+```
 
 ### Configuration
 
-You can configure the gem using the built-in configuration command:
+Configure the API URL and model interactively:
 
 ```bash
 ai-commit-message config
 ```
 
-This interactive prompt allows you to set:
-- **API URL**: The endpoint for your AI model (defaults to `http://localhost:11434`)
-- **Model name**: The AI model to use (defaults to `qwen2.5-coder:7b`, which I personally recommend)
+If your server is reachable, the model prompt becomes a picker populated with the models available on that server.
+
+List the models available on your server:
+
+```bash
+ai-commit-message models
+```
 
 ### Command Line Options
 
@@ -64,44 +87,59 @@ Override configuration settings directly:
 ai-commit-message commit --url=http://your-api-endpoint --model=your-model-name
 ```
 
+All options:
+
+| Option | Description | Default |
+| --- | --- | --- |
+| `--url` | Base URL of the OpenAI-compatible API | `http://localhost:11434` |
+| `--model` | Model name | `qwen3:8b` |
+| `--length` | Maximum message length in characters | `72` |
+| `--conventional` | Use Conventional Commits format | off |
+| `--message-only` | Print the message without the commit confirmation prompt | off |
+
 ## How It Works
 
 The gem:
-1. Collects your staged changes using `git diff --cached`
+1. Collects your staged changes using `git diff --cached` (capped at 12,000 characters so large diffs don't overwhelm a local model's context)
 2. Retrieves your recent commit history for context
 3. Identifies your current branch name
-4. Sends this information to the specified AI model
-5. Returns a concise, contextual commit message (limited to 250 characters)
+4. Sends a chat request to `{url}/v1/chat/completions` — the standard OpenAI-compatible endpoint
+5. Cleans the model output (strips code fences, quotes and reasoning blocks) and prints a single-line commit message (limited to 72 characters by default)
+
+The URL accepts base addresses with or without a trailing `/v1` and `/`, so `http://localhost:11434` (Ollama), `http://localhost:1234/v1` (LM Studio) and `http://localhost:8080` (llama.cpp `llama-server`) all work as-is.
 
 ## Configuration File
 
 The gem stores your configuration in `~/.ai-commit-message.conf`. You can manually edit this file if needed. Example:
 
 ```bash
 url=http://localhost:11434
-model=qwen2.5-coder:7b
+model=qwen3:8b
 ```
 
 ## Models
 
-The default configuration uses Ollama with the `qwen2.5-coder:7b` model, but you can use any model available through your API endpoint.
+The default configuration uses Ollama with the `qwen3:8b` model, but you can use any model available through your API endpoint.
 
 ### Recommended Models
 
-- **qwen2.5-coder:7b**: Good balance of quality and speed (default)
-- **llama3:8b**: Excellent for general commit messages
-- **codellama:7b**: Specialized for code-related commits
+- **qwen3:8b**: Good balance of quality and speed (default). Note that it is a thinking model, so expect a few seconds of reasoning before the message appears
+- **qwen2.5-coder:7b**: Faster, non-thinking coding model
+- **qwen3.5:4b**: Current-generation small model for modest hardware
+- **gemma4:e4b**: Compact general-purpose option
+- **llama3.1:8b**: Solid general-purpose choice
 
 ## Contributing
 
 1. Fork the repository
 2. Create your feature branch: `git checkout -b my-new-feature`
 3. Install dependencies: `bundle install`
-4. Make your changes and add tests if applicable
-5. Commit your changes: `git commit -m 'Add some feature'`
-6. Push to the branch: `git push origin my-new-feature`
-7. Submit a pull request
+4. Run the tests: `bundle exec rake test`
+5. Make your changes and add tests if applicable
+6. Commit your changes: `git commit -m 'Add some feature'`
+7. Push to the branch: `git push origin my-new-feature`
+8. Submit a pull request
 
 ## License
 
-This gem is available as open source free to use/alter without restrictions.
+This gem is available as open source under the terms of the [MIT License](LICENSE).

Rakefile

@@ -0,0 +1,9 @@
+require 'rake/testtask'
+
+Rake::TestTask.new do |t|
+  t.libs << 'lib' << 'test'
+  t.test_files = FileList['test/**/*_test.rb']
+  t.warning = false
+end
+
+task default: :test

ai-commit-message.gemspec

@@ -1,21 +1,24 @@
 Gem::Specification.new do |gem|
   gem.name          = 'ai-commit-message'
-  gem.version       = '0.1.2'
+  gem.version       = '0.2.0'
   gem.authors       = ['Afonso Neto']
   gem.email         = ['afonso.pontesneto@gmail.com']
 
   gem.summary       = 'Git commit messages suggested by AI'
-  gem.description   = 'Git commit messages suggested by AI'
+  gem.description   = 'Git commit messages suggested by AI using a local LLM via any OpenAI-compatible API'
   gem.homepage      = 'https://github.com/AfonsoNeto/ai-commit-message'
-  gem.license       = nil
-
-  gem.files         = Dir.glob('{bin/*,lib/**/*,README.md}')
+  gem.license       = 'MIT'
+
+  gem.metadata['rubygems_mfa_required'] = 'true'
+
+  gem.files         = Dir.glob('{bin/*,lib/**/*,README.md,LICENSE}')
   gem.require_paths = ['lib']
   gem.executables   = ['ai-commit-message']
   gem.required_ruby_version = '>= 3.0'
 
   gem.add_dependency 'thor'
-  gem.add_runtime_dependency 'tty-prompt'
-  gem.add_runtime_dependency 'net-http'
-  gem.add_runtime_dependency 'json'
-end
+  gem.add_dependency 'tty-prompt'
+
+  gem.add_development_dependency 'rake'
+  gem.add_development_dependency 'minitest'
+end

docs/demo.cast

@@ -0,0 +1,39 @@
+{"version": 3, "term": {"cols": 84, "rows": 13}, "timestamp": 1790494557, "command": "ai-commit-message"}
+[0.8, "o", ""]
+[0.84, "o", "\u001b[2m# ai-commit-message: AI commit messages from a local LLM\u001b[0m\r\n\r\n"]
+[0.89, "o", "\u001b[1;32m$\u001b[0m "]
+[0.94, "o", "ai-c"]
+[0.99, "o", "ommi"]
+[1.04, "o", "t-me"]
+[1.09, "o", "ssag"]
+[1.14, "o", "e co"]
+[1.19, "o", "mmit"]
+[1.24, "o", " --m"]
+[1.29, "o", "essa"]
+[1.34, "o", "ge-o"]
+[1.46, "o", "nly"]
+[2.36, "o", "\r\n"]
+[2.96, "o", ""]
+[3.0, "o", "docs: add type check and docstring to add function\r\n\r\n"]
+[3.05, "o", "\u001b[1;32m$\u001b[0m "]
+[3.1, "o", "ai-c"]
+[3.15, "o", "ommi"]
+[3.2, "o", "t-me"]
+[3.25, "o", "ssag"]
+[3.3, "o", "e co"]
+[3.42, "o", "mmit"]
+[4.32, "o", "\r\n"]
+[4.72, "o", ""]
+[5.22, "o", "docs: add type check and docstring to add function\r\n"]
+[5.62, "o", "Commit with this message? (\u001b[1mY\u001b[0m/n) "]
+[6.22, "o", "y\r\n"]
+[6.26, "o", "\u001b[32m3f7159a docs: add docstring and type check for add function\u001b[0m\r\n\r\n"]
+[6.31, "o", "\u001b[1;32m$\u001b[0m "]
+[6.36, "o", "git "]
+[6.41, "o", "log "]
+[6.46, "o", "--on"]
+[6.51, "o", "elin"]
+[6.63, "o", "e -2"]
+[7.03, "o", "\r\n"]
+[7.53, "o", ""]
+[7.53, "o", "\u001b[33m3f7159a docs: add docstring and type check for add function\r\nb445970 chore: initial setup\u001b[0m\r\n\r\n"]