v2.0.0 - Vision Command Redesign

Breaking Changes

• Vision command now requires -f flag for image input instead of positional argument
  • Old: syn vision image.jpg "prompt"
  • New: syn vision -f image.jpg "prompt"
  • This change improves CLI consistency with the chat command's -f file flag

Documentation

• Updated README vision examples to reflect new -f flag syntax
• Regenerated usage screenshot with updated help text
• Updated model example from qwen to coder for variety

Commits

• f62f00e feat(vision): use -f flag for image input instead of positional arg
• 751ae0a docs: update vision examples to use -f flag
• 77736d1 docs(assets): regenerate usage screenshot with updated help text

v1.2.0 - Model List Redesign

What's Changed

Features

• Redesigned model list output - Compact single-line format with copy-paste ready hf: IDs, dynamic column alignment, inline alias tags, vision capability indicators, and alphabetical sorting for easier scanning

Full Changelog: v1.1.0...v1.2.0

v1.1.0 - Model Updates

What's Changed

Model Updates

• Updated model aliases: kimi now points to K2.5 (from K2-Thinking), qwen now points to Qwen3-235B-Thinking (from VL-235B vision model)
• New model aliases: Added coder (Qwen3-Coder-480B), r1 (DeepSeek-R1), minimax (MiniMax-M2.1), llama (Llama-3.3-70B)
• Vision command fix: Switched default vision model from qwen to kimi (the only remaining vision-capable model)

Documentation

• Updated all documentation to reflect new model aliases
• Updated help text and CLI flags with complete model list

Full Changelog: v1.0.0...v1.1.0