# Pastea [Health: Active]

**Category:** 🖥️ OS Automation  
**Repository:** https://github.com/pastea-app/pastea-mcp  
**GitHub Stars:** 0  
**npm Downloads (last month):** 109  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/pastea

## Description
Search, read and re-copy your Mac clipboard history from Pastea. Local, pairs with one click.

## Tools
Capabilities this server exposes over MCP:

- **pastea_status** — Reports whether Pastea is reachable and connected on this Mac, and what to do if it is not.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "pastea": {
    "command": "npx",
    "args": ["-y","@pastea/mcp"]
  }
}
```

## Documentation

## What Pastea MCP server does

Pastea MCP server connects AI applications to clipboard history managed by Pastea on a Mac. It lets an assistant find earlier copies of text, links, file names, and text recognized inside screenshots. The assistant can retrieve a selected clip, view an image thumbnail, list available history collections, and place a clip back on the system clipboard.

The bridge does not send clipboard data to a third-party service. It communicates with Pastea on the local machine and passes data to the AI application that launched it after access is approved. Pastea-concealed items, such as clips hidden by password managers, are not exposed.

## How it works

The package is a dependency-free bridge launched by an MCP client over stdio. Pastea itself provides the MCP endpoint on 127.0.0.1. On the first connection, Pastea displays an approval request for the specific AI application. Selecting Allow grants access; selecting Don't Allow leaves the connection without clipboard access and exposes status information explaining the problem.

Pastea MCP server stores the issued access token in `~/Library/Application Support/Pastea/MCP Bridge/tokens.json`. The token is readable by the local user account, and revoking the connection in Pastea invalidates it. Connections can be reviewed or revoked under Pastea's MCP and AI Tools settings.

## Setup and configuration

The supported requirements are macOS 15 or newer, Pastea 1.3 or newer with Pastea Pro, and MCP enabled in Pastea's Settings → MCP & AI Tools. Node.js 20 or newer is needed for installations using `npx`; Claude Desktop's packaged installation does not require a separate Node.js installation.

Claude Desktop can be connected from Pastea's settings, through Claude Desktop's extension browser, or with the release `.mcpb` package. Other stdio-capable applications can use:

```bash
npx add-mcp @pastea/mcp
```

Individual configurations use `npx -y @pastea/mcp`. Claude Code and Codex have dedicated `mcp add` commands, while Cursor and VS Code provide installation links. If tools do not appear, fully restart the AI application so it relaunches the bridge. The optional `PASTEA_MCP_URL` setting can point a development run at a specific local Pastea endpoint.

## Tools and capabilities

Pastea MCP server exposes these tools:

- `search_clips` searches clipboard history and user-created lists.
- `get_clip` reads a complete clip and can provide a picture thumbnail.
- `list_collections` returns history and lists with item counts.
- `copy_clip` restores a clip to the system clipboard; the user performs the final paste with Command-V.
- `pastea_status` reports whether Pastea is reachable and connected and indicates how to resolve connection issues.

## Limitations and notes

The assistant cannot delete clips or paste directly into another application. It cannot access concealed clips. Pastea MCP server only operates after the user approves the connection, and access may need to be granted again after a revocation or Pastea reset. Changing Pastea's port is supported because the bridge reads the port from Pastea preferences whenever it starts; the AI application must be restarted afterward.

_Full upstream README: https://allmcps.com/mcp/pastea/readme_