Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces projectmem, a local-first memory and judgment layer designed to enhance AI coding agents by maintaining a persistent, event-sourced log of development activities. By capturing decisions, attempts, and fixes in an append-only format, projectmem reduces the need for repetitive context reconstruction and prevents agents from repeating previously failed actions. Evaluated over two months across 10 projects, the system demonstrates significant improvements in efficiency and decision-making, serving as a foundational tool for reproducible AI-assisted development.
AI coding agents can now avoid costly debugging cycles by leveraging a persistent memory system that remembers past decisions and actions.
AI coding assistants now support a growing share of software work, from quick scripts to production applications. Yet these agents remain largely stateless: each new session re-reads project files, re-derives prior decisions, and - most costly - may repeat debugging attempts that already failed. Reconstructing this context can consume an estimated 5,000-20,000 tokens per session; the bottleneck is often not model capability but missing project memory. We present projectmem, an open-source, local-first memory and judgment layer for AI coding agents. projectmem records development as an append-only, plain-text event log of typed events - issues, attempts, fixes, decisions, and notes - and deterministically projects that log into compact, AI-readable summaries served through the Model Context Protocol (MCP). Beyond storage, projectmem adds a deterministic pre-action gate that warns an agent before it repeats a previously failed fix or edits a known-fragile file. We frame this as Memory-as-Governance: memory that does not merely answer the agent but acts on its next action. The system runs fully offline with no telemetry; its immutable log also serves as a provenance trail for reproducible, auditable AI-assisted development. projectmem ships as a three-dependency Python package (14 MCP tools, 19 CLI commands, 37 automated tests) and is evaluated through a two-month self-study across 10 projects comprising 207 logged events. Source code: https://github.com/riponcm/projectmem.