A multi-turn CLI chatbot that handles context window limits with two real strategies — sliding window and summarisation compression — and shows you exactly what each approach loses. Every multi-turn AI product hits this problem eventually; this project makes you the person on the team who already knows the tradeoffs instead of discovering them in production.
Every production chatbot eventually hits this wall — being able to name and implement the tradeoff is what separates a demo-builder from someone who can ship a long-running assistant.