v3.3.0

What's New

✨ Print LLM requests & responses to stdout (print_stream) — #78

New print_stream config option prints every LLM request and response to stdout, with streamed responses appearing token by token as they are generated. Works for both streaming and non-streaming client requests. Intended for local development and troubleshooting.

print_stream = true

See Printing LLM Requests and Responses in the README.

🐛 Bug Fixes

• Config files with UTF-8 BOM are now loaded correctly (TOML/JSON saved by some Windows editors previously failed to parse) — #80
• Prompts printed via print_stream are no longer truncated — #78
• Versioned Docker images are now published on every GitHub release (3.3.0, 3.3, 3, latest); previously only edge was built — #81, closes #70

🔧 Maintenance

• Updated ai-microcore to 6.6.1

Full Changelog: v3.2.3...v3.3.0