session-engram 1.0.1

Cross-session memory system for AI coding assistants. Persists conversation context, task progress, and reusable experiences across sessions via Markdown files. Features Hook-powered auto-recall, interactive knowledge graph visualization, chronological timeline, and global experience library. Supports Claude Code, Cursor, and OpenCode.

Session-Engram

Engram: The physical trace of memory in the brain.

Cross-dialogue session memory system — persistent context, reusable experience, AI automatic recall.

English | 中文

---

What is Session-Engram?

Session-Engram is a cross-session memory system for AI coding assistants. It solves two fundamental problems:

1. Cross-Session Memory Loss — When a conversation gets too long, starting a new session yields better results. But the new session cannot reuse the previous session's memory.

2. Experience Doesn't Carry Over — Effective solutions and lessons learned during programming live only in the current session and are lost when it ends.

How It Works

Session-Engram stores session memory and experiences as structured Markdown files, then uses a Hook mechanism to make the AI automatically read them at the start of each new session.

• Storage layer: Sessions and Experiences stored as Markdown files — human-readable, AI-readable, Git-trackable
• Index layer: index.md is a lightweight summary, avoiding stuffing all memory into context (saves tokens)
• Activation layer: Hook + rule injection makes AI "passively triggered" rather than "actively remembering"

Quick Start

pip install session-engram

# Initialize .engram directory (auto-installs AI platform integration)
sengram init

# Generate memory map (relationship graph)
sengram map

# Generate timeline (log view)
sengram timeline

# Generate memory index (AI entry point)
sengram index

Documentation

• Full English Documentation: README_EN.md
• 完整中文文档: README_CN.md

License

MIT - See LICENSE