Last released Oct 1, 2026
Fit images to a vision-LLM token budget before you send them (Python bindings for vision-squeezer)