AI Release

airelease creates GitHub releases from merged pull requests and can ask OpenRouter (or custom endpoints) to suggest a semantic version and release title options.

Features

• Interactive release creation flow
• Auto-discovers PRs since the latest release
• Uses gh release create to publish
• Global config in ~/.airelease/config.json
• Per-repo configurable base branch
• Configurable OpenRouter model or custom endpoint
• Dry run mode to preview changes

Requirements

• git
• gh (authenticated with gh auth login)
• Go 1.22+

Install

From this repository:

go install .

Ensure your Go bin path is in PATH:

export PATH="$(go env GOPATH)/bin:$PATH"

Quick Start

Using OpenRouter

1. Set your OpenRouter API key:

airelease config openrouter-api-key <your-api-key>

2. Optional: set model globally:

airelease config model qwen/qwen3.5-flash-02-23

3. Optional: set base branch for current repo:

airelease config base develop

4. Run it inside a git repository:

airelease

Using a Custom Endpoint

1. Set your custom endpoint:

airelease config model-url "https://your-endpoint.com/v1"

2. Add required headers (e.g., Authorization):

airelease config model-header "Authorization: Bearer your-token"

3. Set your model name:

airelease config model "your-model-name"

4. Run it:

airelease

Dry Run Mode

Test without creating a release:

airelease --dry-run

Commands

Basic Usage

airelease                    # Create a GitHub release from merged PRs
airelease --dry-run          # Preview release without creating it

Configuration

airelease config base <branch>                    # Default base branch for current repo
airelease config openrouter-api-key <api-key>     # Set OpenRouter API key
airelease config openrouter <true|false>          # Enable/disable OpenRouter mode  
airelease config model <model-name>               # Set model name

Custom LLM Endpoints

airelease config model-url <url>                 # Set custom endpoint (full path)
airelease config model-header <key:value>        # Add custom header
airelease config model-header remove <key>       # Remove a header
airelease config model-header list               # List all custom headers

Note: When using model-url or model-header, OpenRouter mode is automatically disabled. The tool will append /chat/completions to the base URL if not present.

Notes

• If OpenRouter is not configured, airelease still works without AI suggestions.
• API key is stored in ~/.airelease/config.json with file mode 0600.
• Custom endpoints are supported with the model-url and model-header config commands.
• dry-run mode previews the release without actually creating it.