Pastea MCP
OfficialClick on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Pastea MCPsearch my clipboard history for the link I copied earlier"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Pastea MCP
Local MCP server for Pastea. Give Claude, Cursor, Codex, VS Code and other AI tools access to your Mac's clipboard history — without anything leaving your Mac.
Your assistant can search what you copied, pull a clip into context (including the text Pastea reads inside your screenshots), and put a clip back on the clipboard for you to paste.
Pastea runs the MCP server itself, on 127.0.0.1 only. This package is the thin,
dependency-free bridge that AI apps launch: it finds Pastea, asks it for access once —
you click Allow in Pastea — and relays messages over stdio.
Install
Claude Desktop
One click from Pastea: open Settings → MCP & AI Tools, turn on Enable MCP, and click Claude Desktop. Claude Desktop opens with the Pastea extension ready to install.
Or find Pastea in Claude Desktop under Settings → Extensions → Browse extensions, or
download pastea-mcp-<version>.mcpb from the
latest release and double-click it.
The first time Claude uses it, Pastea asks you to allow the connection.
Every other AI app
Use add-mcp to connect Pastea to all your installed
AI apps at once:
npx add-mcp @pastea/mcpOr connect each app yourself:
Claude Code
claude mcp add pastea -- npx -y @pastea/mcpCodex
codex mcp add pastea -- npx -y @pastea/mcpCursor
VS Code
Anything else that runs stdio MCP servers:
{ "mcpServers": { "pastea": { "command": "npx", "args": ["-y", "@pastea/mcp"] } } }The first time an app uses Pastea, Pastea shows "Allow <app> to use your clipboard history?" — click Allow and the tools appear.
Connect from Pastea
Every app above can also be connected from inside Pastea — Settings → MCP & AI Tools, click the app's chip. Cursor and VS Code open an install prompt; Claude Code, Windsurf and Codex get the exact command or config to paste, already carrying an access token, so no pairing prompt is needed.

Related MCP server: clippy
What your assistant can do
Tool | What it does |
| Search your clipboard history and lists — text, links, file names, and text recognized in screenshots. |
| Read one clip in full, with a thumbnail for pictures. |
| List your history and the lists you made, with counts. |
| Put a clip back on the system clipboard, exactly as copied. You paste it yourself with ⌘V. |
| Says whether Pastea is reachable and connected, and what to do if not. |
An AI tool can never delete a clip, never pastes into another app, and never sees items Pastea conceals (from password managers).
How pairing works
The bridge ships with no credentials. On first use it asks Pastea for access and Pastea shows
"Allow Claude Desktop to use your clipboard history?" (or Cursor, Codex, …). Click Allow
and the tools appear a moment later; click Don't Allow and nothing is granted — the assistant
gets a pastea_status tool that explains the situation.
Connections are listed in Pastea → Settings → MCP & AI Tools → Connected AI Tools, where you can Revoke each one. Pastea then asks again only if you ask the assistant to reconnect.
Requirements
macOS 15 or later
Pastea 1.3+ with Pastea Pro, and Enable MCP turned on in Settings → MCP & AI Tools
For
npxinstalls: Node.js 20+ (Claude Desktop needs nothing — it ships its own)
Privacy Policy
Pastea MCP runs locally on your Mac.
It connects only to Pastea on
127.0.0.1and to nothing else. It makes no network requests to Nexo Labs or any third party, and carries no analytics or telemetry.It moves clipboard data only between Pastea and the AI app that runs it, and only after you clicked Allow in Pastea. What that AI app does with clips it reads is governed by that app's own privacy policy.
It stores one thing: the access token Pastea issued when you clicked Allow, in
~/Library/Application Support/Pastea/MCP Bridge/tokens.json, readable by your user account only. Revoking the connection in Pastea invalidates it immediately; uninstalling Pastea removes the file.Concealed clips are never available to it, and it cannot delete anything.
Full policy — what data Pastea processes, how it is stored, sharing, retention, and contact: https://pastea.app/privacy.
Troubleshooting
"Pastea isn't running, or its MCP endpoint is turned off."
Open Pastea and turn on Enable MCP in Settings → MCP & AI Tools, then ask the assistant to
call pastea_status.
"Pastea Pro is required."
MCP access is a Pro feature. Subscribe in Pastea's Settings, then call pastea_status.
"Pastea is asking the user to allow this connection." Look for the Pastea window — it may be behind the AI app — and click Allow.
The tools vanished.
The connection was revoked in Pastea, or Pastea was reset. Ask the assistant to call
pastea_status to pair again.
Tools don't show up after connecting. Fully quit and reopen the AI app so it relaunches the bridge.
Changed Pastea's port? The bridge reads it from Pastea's preferences on every launch; restart the AI app.
Logs for Claude Desktop: ~/Library/Logs/Claude/mcp-server-Pastea.log. The bridge never logs
the token.
Development
npm install
npm test # node --test, no network
npm run build # tsc → dist/
scripts/pack.sh # validate + pack → build/pastea-mcp-<version>.mcpb + SHA256SUMSRun it against a development Pastea by hand:
PASTEA_MCP_URL=http://127.0.0.1:41418/mcp node dist/index.jsthen type JSON-RPC lines on stdin (initialize, notifications/initialized, tools/list).
Releases are cut by tagging v<version> (matching package.json, manifest.json,
server.json and src/version.ts); the workflow packs the bundle and publishes it.
Contributing
Issues and pull requests welcome — see the issue tracker.
License
Available Tools
1 toolpastea_statusPastea StatusARead-only
Reports whether Pastea is reachable and connected on this Mac, and what to do if it is not.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true and openWorldHint=false; the description adds that the check is scoped to 'this Mac' (local, not remote) and that the response includes remediation guidance when the check fails. That is genuine value beyond the structured fields, though it says nothing about response shape or failure modes of the check itself.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single tight sentence that front-loads the core question (reachable/connected) and then the follow-up value (remediation). Nothing is padded or redundant.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a zero-parameter read-only probe with annotations covering safety and no output schema, the description covers what the tool answers and that it includes next-step advice. It stops short of describing what 'connected' means or what the remediation content looks like, a minor gap given the simplicity of the tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool takes zero parameters, so there is nothing for the description to disambiguate; baseline 4 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Specific verb ('Reports') plus a concrete subject ('whether Pastea is reachable and connected on this Mac'), so an agent immediately knows this is a local health/connectivity check. There are no siblings to differentiate from, so the lack of routing language costs nothing.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The phrase 'what to do if it is not' implies the tool should be called when Pastea seems unreachable, which is implied usage rather than an explicit when/when-not statement. With no alternatives existing, there is nothing to route against, but the trigger condition is never spelled out.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
1 tool update
v1.0.4- First observed
pastea_status
TDQS
Scored across 1 tool
With only one tool, there is no possibility of overlap or misselection; pastea_status has a single unambiguous purpose. Distinctness is trivially satisfied rather than earned through careful scoping.
The lone name pastea_status uses clean snake_case and is readable, but it is a noun_status form rather than a verb_noun action pattern, and a single tool provides no pattern to demonstrate consistency against.
One tool is far too thin for a server branded as 'Pastea MCP', which implies paste/content management functionality. A single connectivity probe cannot represent the apparent scope of the service.
The surface exposes only a reachability check, with no create, read, list, update, or delete operations for pastes. An agent has no way to actually do anything with Pastea, making this a severe gap for the stated purpose.
Maintenance
Related MCP Connectors
Give Claude, Cursor, or another AI assistant direct access to your QuickSnip library.
Mac & Windows: let ChatGPT, Claude & Cursor use your email, calendar, iMessage, Teams, files. Free.
Use your Mac, Windows or Linux computer from ChatGPT, Claude or Codex: files, commands, documents.
Search, read, and write your Apple Notes from ChatGPT/Claude via a local Mac agent + MCP relay.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceProvides AI assistants access to the macOS clipboard content, supporting text, images, and binary data via OSAScript.8MIT
- AlicenseNot gradedqualityAmaintenanceGives AI assistants direct access to macOS clipboard - letting Claude copy generated code, text, or files straight to your clipboard for pasting anywhere on Mac.249MIT

ContextPulseofficial
AlicenseAqualityBmaintenanceLets AI assistants understand what you're working on — current screen content, recent dictation, clipboard, and saved notes — running entirely on your own machine with nothing sent to the cloud.36AGPL 3.0- AlicenseNot gradedqualityCmaintenanceEnables Claude to browse, search, and restore clipboard history with type classification, secret detection, and Touch ID-gated decryption.19MIT