# Reduce cost and drift with out-of-band memory optimization (beta)

> A new MemoryOptimizationStrategy framework and APIs allow you to summarize and optimize memories outside of agent runs.

- Published: 2025-11-27
- Author: Agno Team
- Category: Changelog
- Canonical: https://agno-com-nine.vercel.app/articles/reduce-cost-and-drift-with-out-of-band-memory-optimization-beta
- Markdown: https://agno-com-nine.vercel.app/articles/reduce-cost-and-drift-with-out-of-band-memory-optimization-beta.md

A new MemoryOptimizationStrategy framework and APIs allow you to summarize and optimize memories outside of agent runs. By decoupling memory maintenance from inference, you can keep context high-signal while reducing runtime tokens and improving decision quality at scale.

**Details**

- Schedule or trigger memory compaction and summarization independently of agent runs
- Keep knowledge current and concise to improve downstream model performance
- Works without changes to agent logic; designed for scale and governance

**Who this is for:** Production teams with large or fast-growing memory stores seeking lower costs and tighter control.