Lisa – Long Term Memory for AI Coding Assistants

!Lisa for Claude

> Lisa, never forgets a commit, a pr, a code change, or a saxophone lesson.

Why Lisa?

$3

You want..

- memory to load automatically
- store automatically
- recall automatically

You DON'T want..

- Having to repeat yourself, over & over
- Having to update .md files
- Having to fill up your context window.

How?

lisa init in your project directory. Open Claude cli and code away as normal.

$3

- Lisa, tell me about this project?
- Lisa, what was the last thing we were working on?
- Lisa, why are we not using Postgres?
- Lisa, ...

$3

- Lisa, remember this for future reference
- Lisa, create Jira/GitHub ticket
- Lisa, let's commit, push, and create a PR

---

The tech..

- Graph-native storage (Neo4j) - Connections matter as much as content
- LLM-powered extraction - Automatically identifies entities and relationships
- Temporal awareness - Knows when you learned something
- Semantic retrieval - Finds relevant context by meaning, not keywords
- Multi-CLI support - Works with Claude Code and OpenCode

---

!Graph DB

Installation

$3

``bash


Install Lisa globally - IMPORTANT

npm install -g @tonycasey/lisa

change to any project directory

cd your-project-directory

initialize - copies in relevant files

lisa init

``

Using Lisa

Once installed, Lisa works automatically. Your AI assistant will:

1. Load context at session start - Previous memories and project context
2. Capture important info during coding - Decisions, patterns, etc.
3. Remember explicitly when asked - Say "remember that..." to save important notes

During a coding session:

- "remember that we decided to use Redux for state management"
- "hey lisa, what do you know about the authentication system?"
- "lisa, show me recent memories"
- "lisa, what tasks are we working on?"

---

See the Getting Started Guide

---

Contributing | Changelog