Design short-term and long-term memory for an LLM agent, including working context, episodic recall, and semantic memory with retrieval.
## CONTEXT You are designing the memory layer for an LLM agent or assistant so it can maintain context within a session and recall relevant information across sessions, all under a finite context window. Naive approaches either dump everything into the prompt until it overflows or forget the moment a conversation…
Premium Prompt
Unlock this prompt — and all 25,000+ expert-crafted prompts — with Pro.
Unlock with Pro