pax-beehive/paxm

Persistent, provider-neutral memory for Codex, Claude Code, OpenCode, Pi, and MCP coding agents.

What it solves

PAXM is a memory adaptor that prevents users from having to repeatedly explain project context, architecture decisions, and operational constraints to different AI coding agents. It creates a unified memory layer that allows a decision captured in one agent (e.g., Codex) to be recalled in another (e.g., Claude Code), ensuring consistency across different AI tools.

How it works

PAXM acts as an intermediary between AI agents and memory providers. It provides two primary interaction paths:

  • Active Path: Users can explicitly record or recall memories using a CLI, MCP (Model Context Protocol) server, or specific agent skills.
  • Passive Path: Using lifecycle hooks, PAXM automatically recalls relevant context before an agent responds and captures completed conversation turns after the interaction.

It supports multiple interchangeable memory providers, including a built-in SQLite backend for zero-setup local memory, as well as external providers like Zep, Mem0, MemOS, and OpenViking. To ensure the coding session isn't slowed down, passive writes are committed to a local durable queue and delivered to providers in the background.

Who it’s for

Developers who use multiple AI coding agents (such as Cursor, Claude Code, Codex, or Cline) and want a persistent, cross-agent project memory that is either local-first or integrated with enterprise memory systems.

Highlights

  • Cross-Agent Compatibility: Works across a wide variety of coding agents via plugins, extensions, and MCP.
  • Zero-Setup Local Memory: Built-in SQLite support allows for immediate use without API keys or external LLM calls for embeddings.
  • Provider Agnostic: Easily switch between local storage and professional memory services (Zep, Mem0, etc.) without reconfiguring the agents.
  • Non-Blocking Integration: Background queuing for writes and strict timeout budgets for recall ensure that memory retrieval doesn't lag the AI agent's response.
  • Durable Provenance: Tracks the origin and scope (user, agent, session) of every memory.