01Chain-of-thought 'think-aloud' reasoning capabilities
02Real-time streaming of thinking blocks for process transparency
03Pre-built implementation templates for TypeScript and Python
042 GitHub stars
05Configurable token budgets ranging from 1k to 32k+ tokens
06Interleaved thinking support for complex tool-based workflows