DEV Community

Cover image for Stop re-explaining your stack: Persistent local memory for Claude Code and Codex
Manucho IA
Manucho IA

Posted on

Stop re-explaining your stack: Persistent local memory for Claude Code and Codex

Context decay is one of the most frustrating aspects of working with CLI coding agents. The moment a session ends, your LLM treats you like a total stranger.

To solve this without relying on external SaaS or complex vector DBs, I built Wife β€” an open-source, zero-dependency memory layer for Claude Code & OpenAI Codex.

Key Features:
Local & Plaintext: Stores facts in Markdown files you control.

Token Ceiling Management: Automatically evicts low-priority context when hitting limits (1200 tokens for user, 800 for project).

Contradiction Resolution: Updates old information instead of stacking conflicting statements.

Auditability: Easily trace where every learned fact came from.

Quick Start:
Bash
curl -fsSL https://raw.githubusercontent.com/ma-nucho-pro/wife/main/install.sh | bash

Check out the source code and documentation on GitHub.

Top comments (0)