Existing memory systems for embodied agents typically inject retrieved memory as static context at episode start, a paradigm we term Ahead-of-time Monolithic Memory Injection (AMMI). However, this static design quickly becomes misaligned with the agent's evolving state and may degrade lightweight executors below the no-memory baseline. To address this, we propose MemCompiler, which reframes memory utilization as State-Conditioned Memory Compilation. A learned Memory Compiler reads a structured Brief State capturing the agent's current execution state and dynamically selects and compiles only relevant memory into executable guidance. This guidance is delivered through a text channel and a latent Soft-Mem channel that preserves perceptual information not expressible in text. Across Alf World, EmbodiedBench, and ScienceWorld, MemCompiler consistently improves over no-memory across open-source backbones (up to +129%), matches or approaches frontier closed-source systems, and reduces per-step latency by 60%, demonstrating that state-aware memory compilation improves both effectiveness and efficiency.
@article{arxiv.2605.07594,
title = {MemCompiler: Compile, Don't Inject -- State-Conditioned Memory for Embodied Agents},
author = {Xin Ding and Xinrui Wang and Yifan Yang and Hao Wu and Shiqi Jiang and Qianxi Zhang and Liang Mi and Hanxin Zhu and Kun Li and Yunxin Liu and Zhibo Chen and Ting Cao},
journal= {arXiv preprint arXiv:2605.07594},
year = {2026}
}