- Home
- Skills
- Agents & Orchestration
- marm-init
More screenshots
Works with the AI Agents you already use
Marm Init
Guided setup for MARM: local-first shared memory, a knowledge graph, and code indexing across Claude, Cursor, Gemini, and any MCP client
Free
marm-init
Example session with this skill installed
I want to set up marm-mcp's persistent memory for Claude and Cursor on my local machine using Docker.
- Read your context and instructions
- Compiled the marm-init
Setting up MARM via Docker...
- Engine pre-flight: image pulled, marm-memory helper CLI confirmed.
- Transport: HTTP on port 8001 (Docker-managed key, never shown in chat).
- Claude: registered via
claude mcp add. - Cursor: entry written to .cursor/mcp.json.
- Health check: http://localhost:8001/health -> ok
- Console: http://localhost:8002
Setup complete! Your agents now share a persistent memory.
Connects securely to your tools. The creator never sees your data.
About this skill
Persistent AI Memory Made Easy
MARM-Init is a specialized setup engine designed to solve the "goldfish memory" problem in AI agents. While most LLMs lose context between sessions, MARM (Memory Accurate Response Mode) provides a persistent, local-first SQLite backend that stores project history, sessions, and notebooks across your entire development lifecycle. Also includes a local management app, knowledge graph, and project indexing. You can visit GitHub for more info https://github.com/Lyellr88/marm-memory
What it does
This skill acts as an intelligent installer and architect. Instead of manually editing JSON configs or managing Docker containers, you invoke this skill to guide you through a pre-flight hardware check, runtime selection, and multi-agent configuration. It handles the "wiring," so your agent can immediately begin saving and retrieving long-term memories.
Why use this skill?
- Cross-Agent Synchronization: Automatically link Claude, Cursor, VS Code, and Gemini to the same shared memory pool.
- Optimized Context Management: Implements automated re-ranking and pruning to ensure your agent stays within token budgets without losing important project details.
- Privacy First: Everything stays on your hardware. It uses a SQLite WAL backend for fast, deterministic local storage.
- One-Command Infrastructure: Generates API keys, configures transport protocols (HTTP/STDIO), and launches a live management dashboard for you.
The Output
The skill produces a fully configured MCP environment. This includes updated configuration files for your specific IDE or CLI agent, a running MARM server instance (Docker or Python), and a live URL for your memory dashboard.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
- Requires local Docker or Python environment.
- Initial setup requires internet for protocol sync.
- SQLite backend is optimized for single-host performance.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
Trust & safety
Security scan
Security scan passed 2 months ago
- Free with an Agensi account
- Software
- Installer MCPAutomatic Setup MCP
PythonMarm MCPMarm Stdio MCPVs Code Github Copilot Agent MCPGrok Remote MCP
Permissions
No special permissions declared or detected