v3.13.0
What's New
- You can enable the new --50-72 flag or set the GAC_USE_50_72_RULE environment variable to keep commit messages short and tidy.
- When the AI is generating a response, the spinner now shows an estimated token count so you know how much data is being processed.
Improvements
- The AI now uses the latest GLM‑5 model, providing more accurate and relevant answers.
- Setup now asks if you want to enable the 50/72 rule, making configuration simpler.
- Outdated references were removed, resulting in clearer documentation and fewer distractions.
v3.12.1
Improvements
- You can now use the newest versions of the smart models
- Korean instructions are now easier to read
- It is easier to find the grouping feature in the help guide
v3.12.0
What's New
- Upgraded to newer, more capable AI models across multiple providers
Improvements
- Simplified token counting for faster and more reliable performance
- Improved release process for better control over software updates
v3.11.0
What's New
- New commands let you view, edit, import from a file, or clear a custom system prompt directly from the tool
Improvements
- The version command now shows a clear list of what changed in each release, so you can see updates at a glance
- Prompt commands now always display the most current default text when no custom prompt is set
- Simplified the prompt commands by removing an unnecessary option, making the interface cleaner
v3.10.17
Improvements
- Browser sign‑in works more reliably
Bug Fixes
- Fixed an issue where opening the sign‑in page could cause the app to hang; now a clear message appears instead