Claude Supermemory V2: Persistent Memory Plugin for Claude Code

Introduction

Claude has become the go‑to LLM for developers, but every time you finish a task you lose the context. Claude Supermemory V2 solves this by adding a persistent, team‑shared memory layer to Claude Code. The plugin is fully open‑source, easy to install and configure, and lets you store key decisions, project architecture, bugs, and more – all automatically captured when a session ends.

What is Claude Supermemory?

• Persistent memory – Store important turns in your code or conversation history. Claude can pull these back when you revisit the project.
• Team & personal scopes – Separate containers let you share knowledge across a project or keep it private.
• Auto‑capture – The plugin automatically saves conversations at the end of each session.
• Custom triggers – Configure keywords like remember, architecture, bug, fix to identify important turns.
• Integrated workflows – Use the built‑in Claude commands, e.g., /claude-supermemory:index or /claude-supermemory:project-config.

Prerequisites

• An active Claude Code environment.
• A Supermemory account (free tier is sufficient; premium unlocks larger context windows).
• Basic command‑line experience.

Installation Steps

1. Open a terminal in the project where you run Claude Code.
2. Run the marketplace command to add the plugin:

   /plugin marketplace add supermemoryai/claude-supermemory
   

3. Install the plugin:

   /plugin install claude-supermemory
   

Configuration

1. Environment Variable

Set your Supermemory API key. Get it in your console at console.supermemory.ai.

export SUPERMEMORY_CC_API_KEY="sm_..."

export SUPERMEMORY_DEBUG=true

2. Global Settings

Create or edit ~/.supermemory-claude/settings.json:

{
  "maxProfileItems": 5,
  "signalExtraction": true,
  "signalKeywords": ["remember", "architecture", "decision", "bug", "fix"],
  "signalTurnsBefore": 3,
  "includeTools": ["Edit", "Write"]
}

3. Project‑Level Overrides

Create .claude/.supermemory-claude/config.json (or use the /claude-supermemory:project-config command):

{
  "apiKey": "sm_...",
  "repoContainerTag": "my-team-project",
  "signalExtraction": true
}

Using the Plugin

During a session, prefix key phrases with the trigger words (remember, bug, fix). The plugin automatically extracts these into the memory pool. On the next run, call

/supersearch ...

Advanced Customization

• Dynamic signal keywords – Adjust signalKeywords for your own workflows.
• Custom tags – Use repoContainerTag to namespace memories by project or domain.
• Debug mode – Set SUPERMEMORY_DEBUG=true to log captures and search queries for troubleshooting.

Contribution Guide

1. Fork supermemoryai/claude-supermemory.
2. Create a feature branch: git checkout -b add-feature-x.
3. Add tests and documentation.
4. Submit a pull request following the template.

The project uses MIT license, so feel free to integrate the plugin into your commercial or private workflows.

Why It Matters

By preserving context across sessions, Claude Supermemory reduces the cognitive overhead of debugging, onboarding, and iterative design. Teams can collaborate on the same knowledge base, ensuring everyone references the same architectural decisions and code patterns.

Closing Thoughts

Claude Supermemory V2 brings a missing piece to the Claude ecosystem – persistent, shareable memory. It’s simple to set up yet powerful in practice. Dive in today, experiment with the triggers, and watch your development workflow become smarter and more efficient.