memorax-ai/memorax-code
A memory plugin for AI coding that turns engineering experience, repository knowledge, and your way of working into memory that remains useful in future tasks.
What it solves
Coding agents often lack continuity between sessions, losing critical context such as architectural decisions, failed attempts, repository-specific rules, and user preferences. MemoraX Code provides a shared memory layer that allows agents to recall prior engineering knowledge and maintain a consistent understanding of a project across different sessions and tools.
How it works
It integrates as a plugin or hook into various coding agents (such as Cursor, Claude Code, and Codex). The system categorizes memory into four boundaries: Coding Memory (engineering lessons), Repo Memory (architecture and project structure), Personal Memory (user preferences), and Procedure Memory (reusable task steps). It uses background writeback to extract reusable knowledge from completed tasks and injects relevant context into new sessions based on a configured cadence or explicit search.
Who it’s for
Developers using AI coding agents who want to maintain long-term project context and avoid repeating the same mistakes or re-explaining preferences across different sessions.
Highlights
- Cross-Agent Compatibility: Works with Codex, Claude Code, CodeBuddy CLI, WorkBuddy, DeepSeek Harness, OpenCode, Trae, and Cursor.
- Background Memory Writeback: Automatically extracts and stores reusable knowledge from completed interactions without manual input.
- Multi-Layered Memory: Separates project-specific architecture, general engineering lessons, user communication preferences, and repeatable procedures.
- Local Observability: Provides local trace and reconciliation records to monitor retrieval and writeback activity.
Related
- Project
- Project
- Project
- Project
- Project