paper-remote-mcp
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
paper-remote-mcp
Run Paper's MCP server without the desktop app — hosted, or on your own machine.
Connect to a hosted server
claude mcp add --transport http paper https://<your-server>/mcp

How signing in works
Your client opens the consent page → Continue.
Open Paper sign-in — sign in on Paper's own page, as usual.
Paper shows an Opening Paper… page. Copy its address from the address bar (⌘L ⌘C) and paste it back.
Done — your client gets a token and stays connected (it refreshes itself).
No Paper password or emailed code ever touches the server. The link you paste is a one-time code, exchanged with Paper over PKCE — the same flow Paper Desktop uses.
Related MCP server: MultiZen
What's true of the server
Nothing is stored. No database, no token logs. Everything handed out is AES-256-GCM ciphertext under one
SEAL_KEY; the session travels inside the client's tokens (~3 KB).Refreshes are automatic. Access tokens live 1 h, refresh tokens 30 days; every refresh asks Paper you're still signed in and folds in Paper's cookie renewals.
Revoking = signing out in Paper. Takes effect within the hour.
Tools run Paper's own live bundle: a lite engine for most calls, Chromium only for the ~8 layout tools (screenshot, export, computed styles, …).
Paper's meter is mocked in the container:
/meter/balanceanswers full (mcpCap: 10000) and/meter/record-mcpis dropped — the weekly-limit gate never trips.
Host it yourself
Any Docker VM (free; one Paper account per server):
docker build -f Dockerfile.deploy -t paper-remote-mcp .
docker run -d --restart=always -p 127.0.0.1:8081:8080 \
-e SEAL_KEY=$(openssl rand -base64 32) -e PUBLIC_URL=https://paper.example.com \
paper-remote-mcpPut your reverse proxy or tunnel (TLS) in front at paper.example.com. One Paper account per server — for you, or each teammate's own.
Cloudflare (multi-user): needs Workers Paid for Containers.
cd remote && npm install
npx wrangler secret put SEAL_KEY # openssl rand -base64 32
npx wrangler deployArchitecture and contract: docs/remote-runtime.md.
Or run it locally
npx paper-remote-mcp login # one-time sign-in in a browser window
npx paper-remote-mcp doctor # browser ✓ sign-in ✓ tools ✓
claude mcp add paper -- npx -y paper-remote-mcp mcpOther clients, same idea:
Cursor —
~/.cursor/mcp.json:{ "mcpServers": { "paper": { "command": "npx", "args": ["-y", "paper-remote-mcp", "mcp"] } } }Codex —
~/.codex/config.toml:[mcp_servers.paper]withcommand = "npx",args = ["-y", "paper-remote-mcp", "mcp"]VS Code —
.vscode/mcp.json:{ "servers": { "paper": { "type": "stdio", "command": "npx", "args": ["-y", "paper-remote-mcp", "mcp"] } } }
Command | What it does |
| Start the stdio MCP server backed by a headless editor. |
| Sign in, or delete the saved session. |
| Check Node, browser, sign-in, and tool availability. |
| Forward stdio to a running Paper Desktop (no CLI needed). |
Handy flags: --file <id|url> (open a file on start), --env staging, --profile <dir>, --headful.
The local server injects a no-op navigator.modelContext so Paper builds its real tool handlers in a headless Chrome page, then maps tools/list/tools/call onto them. The tool list comes from Paper itself, so it tracks Paper's updates.
Local notes
One file open at a time; switching files costs a page reload.
It uses its own Chrome profile (Chrome locks one profile per process), saved under
~/.paper-mcp/profiles/.It depends on Paper's web internals; if Paper changes them,
doctorwill say so.
Programmatic API
import { createEditorHost } from 'paper-remote-mcp';
const host = await createEditorHost({ fileId: '…' });
await host.start();
const result = await host.handleToolCall('agent-1', 'get_basic_info', {}, { name: 'my-app' });
await host.close();Development
git clone https://github.com/vcashwin/paper-remote-mcp && cd paper-remote-mcp
npm install && node src/cli.js doctor
npm test && npm run typecheck # worker tests; typecheck incl. runtime/
node --test runtime/server.test.js runtime/lite.test.jsCommunity project. Not affiliated with or endorsed by Paper. MIT © vcashwin
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Official Browserless MCP server: browser automation, scraping, screenshots and PDFs for AI agents.
Web scraping, browser automation, screenshots and AI data extraction for AI agents via MCP.
Cloud browser, web extraction, screenshots, and remote CDP tools for AI agents.
Live browser debugging for AI assistants — DOM, console, network via MCP.
Related MCP Servers
- AlicenseBqualityAmaintenanceEnables Claude and other MCP clients to drive a real browser by opening pages, clicking elements, filling form fields, capturing PNG screenshots, and extracting visible text. It runs as a lightweight Node process over stdio with lazy-loaded Playwright for fast startup.52MIT
- FlicenseNot gradedqualityAmaintenanceExposes a localhost MCP endpoint that lets AI agents drive isolated browser profiles — navigating, clicking, typing, extracting, screenshotting, evaluating JS, and managing tabs and cookies — so real authenticated workflows can be completed from MCP clients like Cursor or Claude Desktop. Cookies, login state, fingerprints, and proxies persist per profile across launches and AI sessions.-
- AlicenseNot gradedqualityBmaintenanceLets AI agents design alongside you on a live canvas by reading, writing, styling and inspecting real HTML/CSS designs, capturing pixel-accurate screenshots, exporting assets and PDFs, and managing design tokens, comments and files. It exposes the same 37 tool names, parameters and response shapes as Paper, so existing agent prompts and workflows run against a local canvas over Streamable HTTP or stdio.1MIT
- AlicenseNot gradedqualityAmaintenanceConnects MCP-compatible AI agents to an anti-detection browser server, exposing dozens of tools for navigation, page interaction, search, content extraction, downloads, multi-tab handling, and persistent sessions. It lets agents automate and scrape the web at scale using accessibility snapshots, CSS-selector fallbacks, saved profiles, and batch workflows over stdio or authenticated HTTP transport.581 npmMIT