Skip to main content

MemClaw: The Complete Guide to Project Memory for Claude Code

· 4 min read
Felo Search Tips Buddy
Felo Search Tips Buddy
Operations

Complete guide to MemClaw. All features, setup instructions, best practices, use cases, and advanced workflows for persistent project memory.

MemClaw Complete Guide: Everything You Need to Know About AI Workspace Memory

MemClaw is a workspace memory system for Claude Code. It solves the most common frustration with AI coding assistants: every session starts from zero. This guide covers everything — what MemClaw is, how it works, how to set it up, and how to use it effectively. Get started with MemClaw → memclaw.me

What Is MemClaw?​

MemClaw is a persistent memory layer for Claude Code. It connects via MCP (Model Context Protocol) and maintains a structured workspace for each project. When you start a Claude Code session, MemClaw loads your project context automatically. When you end a session, it saves what you built and decided. The next session picks up exactly where you left off. The core problem it solves: Claude Code has no memory between sessions. Every conversation starts from zero. You spend 10-15 minutes re-explaining your project before you can do any real work. The solution: A persistent workspace that Claude queries at session start and writes to at session end.

How MemClaw Works​

Architecture​

Claude Code Session ↓ (MCP Protocol) MemClaw Server ↓ Workspace Store ├── Tech Stack ├── Architecture Decisions ├── Sprint Goals ├── Session History └── Semantic Index

Session Lifecycle​

Session start:

  1. You run claude in your project directory
  2. Claude connects to MemClaw via MCP
  3. Claude queries the workspace for relevant context
  4. Context is loaded into the session
  5. You start working with full project context During session:
  • Claude uses workspace context for accurate suggestions
  • You can query the workspace: "What do we know about authentication?"
  • You can log decisions: "Add to workspace: we decided to use X" Session end:
  1. You run /end
  2. Claude summarizes what was done and decided
  3. Summary is written to the workspace
  4. Next session picks up from here

What Gets Stored​

A MemClaw workspace holds:

Setup​

Step 1: Create a workspace​

Go to memclaw.me, create an account, and create a workspace for your project.

Step 2: Configure MCP​

Create .claude/mcp_config.json in your project root: { "mcpServers": { "memclaw": { "command": "npx", "args": ["-y", "@memclaw/mcp-server"], "env": { "MEMCLAW_API_KEY": "your_api_key", "MEMCLAW_WORKSPACE_ID": "your_workspace_id" } } } }

Step 3: Seed the workspace​

Add your project context to the workspace. Start with tech stack, architecture rules, current sprint goals, and known issues.

Step 4: Create slash commands​

.claude/commands/start.md: Load project context from the MemClaw workspace. Summarize: current sprint goals, last session work, any blockers. Ask what we're working on today. .claude/commands/end.md: Summarize today's session: what was completed, decisions made and why. Save to the MemClaw workspace with today's date. List unresolved issues and suggest next steps.

Step 5: Verify​

Start Claude Code and ask: "What do you know about this project?" If Claude summarizes your workspace content, you're set up correctly.

Multi-Project Setup​

Each project gets its own workspace: ~/projects/ project-a/.claude/mcp_config.json ← MEMCLAW_WORKSPACE_ID=ws_project_a project-b/.claude/mcp_config.json ← MEMCLAW_WORKSPACE_ID=ws_project_b cd into a project, Claude loads that project's context automatically.

Team Setup​

Share a workspace across your team:

  1. One team member creates the workspace and invites others
  2. Each developer uses the same MEMCLAW_WORKSPACE_ID in their config
  3. All sessions load the same context
  4. All session summaries go to the same workspace The workspace becomes a living knowledge base for the entire team.

Best Practices​

Keep it current: Update the workspace when you make decisions. Stale context is worse than no context. Use /end consistently: The session summary is how the workspace grows. Skip it and you lose history. Log decisions immediately: Don't wait until the end of the session. Mid-session decisions are easy to forget. Prune periodically: Remove completed sprint goals and resolved issues. A focused workspace is more useful than a comprehensive one.

Pricing​

MemClaw has a free tier. Paid plans add more workspace storage, team features, and longer history retention. Check memclaw.me for current pricing.

Getting Started​

The setup takes about 20 minutes:

  1. Create account and workspace at memclaw.me
  2. Add MCP config to your project
  3. Seed the workspace
  4. Create slash commands
  5. Run your first memory-enabled session Get started with MemClaw → memclaw.me