# GraphyLoop

> GraphyLoop is an open-source MCP server for developer tools. Workflow planner, wave dispatcher and persistent memory for AI coding agents. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI
- **Website:** https://chadixearth.github.io/graphyloop/
- **Repository:** https://github.com/chadixearth/graphyloop
- **MCP Registry name:** io.github.chadixearth/graphyloop
- **Categories:** developer tools, knowledge and memory
- **Listing updated:** 2026-11-14

## About

GraphyLoop packages a shared working method, called Workflow v3, for coding agents. Its rules are inline-first, one parallel wave at a time, and proof before done, and the same 12 rules apply in every supported harness. Agents are told to run memory_search before planning and memory_store afterwards, saving a dense line with the reason and the commit or PR it landed in.

The npm package wires up seven harnesses in one command: Claude Code, OpenCode, Codex, Cursor and Windsurf, Gemini CLI, Oh My Pi and DeepSeek Harness. Depending on the harness it adds the MCP server, a squad of 25 or 26 role agents, 15 slash commands and 70 bundled Agent Skills. The site lists 15 MCP tools; the registry entry runs it over stdio with npx -y graphyloop mcp.

It is MIT-licensed, has no dependencies and needs Node.js 20 or newer. Claude Code users should install either the plugin or the npx wiring, not both, to avoid duplicate tools.

## Features

- **Persistent memory:** memory_search before planning and memory_store after, with reasoning and commit references.
- **Parallel wave planning:** Plan work as one parallel wave with proof required before done.
- **Role agent squad:** A squad of 25 agents plus bundled Agent Skills.
- **Seven harnesses:** One command configures Claude Code, Codex, Cursor, Gemini CLI and others.
- **Doctor check:** npx graphyloop doctor verifies the wiring.

## Good for

- Give several coding agents the same planning rules
- Keep decisions and their reasons across sessions
- Split a task into a parallel wave of sub-agents
- Run a deploy preflight before shipping

## Install

### Claude Code (Terminal)

```sh
claude mcp add graphyloop -- npx -y graphyloop
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "graphyloop": {
      "command": "npx",
      "args": [
        "-y",
        "graphyloop"
      ]
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "graphyloop": {
      "command": "npx",
      "args": [
        "-y",
        "graphyloop"
      ]
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "graphyloop": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "graphyloop"
      ]
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "graphyloop": {
      "command": "npx",
      "args": [
        "-y",
        "graphyloop"
      ]
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "graphyloop": {
      "command": "npx",
      "args": [
        "-y",
        "graphyloop"
      ]
    }
  }
}
```

## Tools

- `memory_search`: Search stored memory before planning.
- `memory_store`: Store a dense memory line after finishing work.

## FAQ

### Which harnesses are supported?

Claude Code, OpenCode, Codex, Cursor and Windsurf, Gemini CLI, Oh My Pi and DeepSeek Harness.

### What licence is it?

MIT.

### What does it need?

Node.js 20 or newer; it has zero dependencies.

### Should I use the plugin and npx together?

No, use one or the other for Claude Code.

### Is GraphyLoop free?

Yes. GraphyLoop is open source, so you can use and self-host it for free.

### How do I add GraphyLoop to Claude Code?

Run this command in your terminal: claude mcp add graphyloop -- npx -y graphyloop

### What are the best alternatives to GraphyLoop?

The closest alternatives to GraphyLoop are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Ouroboros](https://ai-powered.com/mcp/ouroboros): Ouroboros is an open-source MCP server for developer tools.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Nautilus Compass Memory](https://ai-powered.com/mcp/nautilus-compass-memory): Nautilus Compass Memory is an open-source MCP server for knowledge and memory.

---
Source: https://ai-powered.com/mcp/graphyloop
