Extract reusable coding-agent memory from past sessions with Lerim

Watch Claude Code, Codex CLI, Cursor, or OpenCode sessions, extract durable project memory, and keep it locally as reusable markdown.

Prerequisites

Python 3.10+, Docker recommended

Installation

Use the upstream install or setup path that matches your environment:

• pip install lerim
• uv venv && source .venv/bin/activate
• uv pip install -e '.[test]'

Requirements and caveats from upstream:

• <a href="https://pypi.org/project/lerim/"><img src="https://img.shields.io/pypi/pyversions/lerim?style=flat-square" alt="Python versions"></a>
• python clean_to_lerim_jsonl.py \

Basic usage or getting-started notes:

• Instead of replaying raw traces or losing what happened after each run, Lerim keeps:

• Every run leaves a trace. Most traces are too long, too noisy, and too platform-specific for the next agent to reuse directly.

• bash

• Source: https://github.com/lerim-dev/lerim-cli

• Extracted from upstream docs: https://raw.githubusercontent.com/lerim-dev/lerim-cli/HEAD/README.md

Documentation

• https://github.com/lerim-dev/lerim-cli

Source

• Agent Skill Exchange