Search papers, labs, and topics across Lattice.
This paper systematically explores filesystem-based memory for LLM agents, focusing on how agents can manage, search, and execute tasks using a memory structured as a directory tree of markdown files. The study reveals that while organized memory can significantly reduce retrieval costs, the effectiveness of organization diminishes unless a strong management agent is employed, and no agent effectively translates organization into improved answer quality. Additionally, the findings highlight that altering the toolset can have as much impact on memory structure as changing the underlying model, suggesting a need to rethink memory design in LLMs.
Organized memory can halve retrieval costs, but without a strong management agent, LLMs struggle to maintain effective organization and answer quality.
Deployed LLM agents increasingly keep their long-term memory as a filesystem: a directory tree of markdown files that the agent itself reads, writes, and reorganizes through generic file tools. Yet research has largely passed over this medium: prior systems design bespoke memory representations and study retrieval over them, leaving the default's two working assumptions untested: that an agent can keep a growing store organized as memories accumulate, conflict, and go stale, and that this organization pays. We present the first systematic exploration of filesystem-based memory for LLM agents. We formalize the setting as three roles around one memory filesystem: a management agent integrates and organizes incoming content, a search agent answers queries with cited sources, and an execution agent supplies task trajectories that are distilled into skills, unifying declarative memory and skills in a single store. Across long-conversation benchmarks and embodied tasks, we vary memory shape (agent-organized hierarchy, verbatim dump, chunk retrieval), stream scale, tool harness (sandboxed shell, memory-tool-style functions, varied search tooling), and the strengths of the management and search agents, tracking answer quality, cost, and store health as memory grows. What organization reliably buys is search economy: organized stores roughly halve retrieval cost where material is large. Today's agents, however, fall short of the default's promise: in our growth study, organization erodes for all but the strongest management agent, and no agent we measure converts organization itself into better answers. And the model is not the only lever over a store's shape: changing the tool set alone reshapes the store as strongly as swapping the model. The study turns the filesystem default from an assumption into a design space for agent memory.
