Adding Memory to Multi-Agent Systems

mem0-multica

Add persistent memory to multi-agent systems using Mem0 and Multica. Each agent (Claude Code, Codex, Gemini) gets its own scoped memory via --agent-id — no API key needed, just your email.

Tutorials

Platform	Description	Link
Multica	Give each Multica agent (Claude Code, Codex, Gemini) its own scoped memory via --agent-id	Adding Memory to Multi-Agent Systems — mem0-multica/

Quick Start

# Install mem0 CLI
pip install mem0-cli

# Login with email
mem0 init --email you@example.com

# Pick a tutorial
cd "Adding Memory to Multi-Agent Systems — mem0-multica/"
make setup EMAIL=you@example.com
make demo

How It Works

┌──────────────────────────────┐
│    Multi-Agent Platform      │
│  (Multica, CrewAI, etc.)     │
│                              │
│  Agent A    Agent B    ...   │
│     │          │             │
│     └────┬─────┘             │
│          │                   │
│     ┌────▼─────┐             │
│     │ Mem0 CLI │             │
│     │ per-agent│             │
│     │  memory  │             │
│     └──────────┘             │
└──────────────────────────────┘

Each agent gets its own memory lane via --agent-id. Before a task, search for what the agent knows. After a task, store what it learned.

Adding a New Tutorial

Create a folder for your platform with:

• README.md — setup + usage guide
• Makefile — make setup, make demo, make add, make search
• scripts/ — automation scripts
• examples/ — runnable demos

License

Apache 2.0