# AIMEAT

An MCP server that keeps your notes, agents and apps on a node you control, then serves them to whichever AI you already use.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 10
- Author: miikkij
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/miikkij/aimeat-protocol
- Updated: 2026-10-23
- HTML: https://affpapa.org/en/nevibecode/mcp/aimeat

## What it does

AIMEAT stands for AI Memory Exchange and Action Transfer, and the point is continuity. Ask Claude or Cursor to remember a decision and the record survives the next conversation and shows up in the next tool. Sharing works too, with people and agents you pick. Your AI can also put together a small app, hand a job to an agent or fire a task on a schedule, and every action stays attached to your name.

Two routes in. Point a connector at `https://aimeat.io/v1/mcp`, sign in, then pick the permission level: read only, standard, full, or a custom set. Or run your own node with `npx aimeat init` and `npx aimeat start`; that needs Node.js 24 or later. SQLite works without a database server, PostgreSQL is the production choice. Opening a node to the internet means working through the deployment checklist first, because a fresh install is not ready for a public port.

For a terminal-free setup there is AIMEAT Personal Node on Windows, Apple-silicon Macs and Linux. It wires up the AI tools on the machine and can host a node on port 41050. The installers are not code-signed yet, so both Windows and macOS will ask you to confirm the first launch.

## Requirements
- Remote server: `https://aimeat.io/v1/mcp`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http aimeat https://aimeat.io/v1/mcp
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add aimeat --url https://aimeat.io/v1/mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "aimeat": {
            "url": "https://aimeat.io/v1/mcp"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "aimeat": {
            "type": "http",
            "url": "https://aimeat.io/v1/mcp"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http aimeat https://aimeat.io/v1/mcp
```

## When to use it
- keep context from one chat available in the next tool
- attach Claude or Cursor to memory you host yourself
- run a personal node without sending data to a third party
- turn a description into a small app published on the node
- cut off an AI client's access when you stop trusting it

## Example prompts
- Connect the AIMEAT MCP server at `https://aimeat.io/v1/mcp` and tell me what you can do there
- Remember this choice through AIMEAT so I can find it next week
- Build the app I just described and publish it on my node

## Similar
- [artel](https://affpapa.org/en/nevibecode/plugins/artel.md) — plugin · ★ 9
- [Synaplan](https://affpapa.org/en/nevibecode/mcp/synaplan.md) — MCP server · ★ 287
- [Montycat MCP - Shared Memory for AI Agents](https://affpapa.org/en/nevibecode/mcp/montycat-mcp-shared-memory-for-ai-agents.md) — MCP server · ★ 23
- [strata](https://affpapa.org/en/nevibecode/mcp/agenerationforwordz-tech-strata.md) — MCP server · ★ 6
- [Semurg agent memory](https://affpapa.org/en/nevibecode/mcp/semurg-agent-memory.md) — MCP server
- [Feedbacks](https://affpapa.org/en/nevibecode/mcp/feedbacks.md) — MCP server · ★ 36

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md