About shortcuts-bridge
shortcuts-bridge is an MCP connector for ChatGPT users who want to find and run Apple Shortcuts. A local Mac agent lists and runs shortcuts, with an authenticated relay carrying requests between ChatGPT and the Mac.
Key features
- 01List shortcuts and folders
- 02Search shortcuts by keyword
- 03Run shortcuts with optional text input
- 04Return shortcut output to ChatGPT
Use cases
- 01Find an existing shortcut by topic
- 02Run a morning briefing shortcut
- 03Trigger a Mac automation from chat
MCP support
Exposes shortcut listing, search, folder listing, and execution as MCP tools for ChatGPT. A paired Mac agent performs the actions locally; ChatGPT confirms before running a shortcut.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://shortcutsbridge.vercel.app/mcpclaude mcp add --transport http isaiahdupree-shortcuts-bridge https://shortcutsbridge.vercel.app/mcp{"mcpServers":{"isaiahdupree-shortcuts-bridge":{"type":"http","url":"https://shortcutsbridge.vercel.app/mcp"}}}Open in Cursor{"servers":{"isaiahdupree-shortcuts-bridge":{"type":"http","url":"https://shortcutsbridge.vercel.app/mcp"}}}Open in VS Code{"mcpServers":{"isaiahdupree-shortcuts-bridge":{"type":"http","url":"https://shortcutsbridge.vercel.app/mcp"}}}{"mcpServers":{"isaiahdupree-shortcuts-bridge":{"serverUrl":"https://shortcutsbridge.vercel.app/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEShortcutsBridge is a Model Context Protocol (MCP) connector that lets ChatGPT list, search, and run your Apple Shortcuts. Reads execute locally and requested results pass through the authenticated relay — every action runs on your own Mac through a tiny local agent. The cloud piece is only a stateless relay that shuttles requests between ChatGPT and your Mac.
ChatGPT ──OAuth──► ShortcutsBridge relay ──job queue──► shortcuts-agent ──► Shortcuts.app
(connector) (Vercel, stateless) (your Mac, polls & executes) (via the shortcuts CLI)
Your Mac is the only place your shortcuts are ever listed or run. The relay stores only short-lived job payloads and your account record; it never sees a shortcut unless a job is in flight, and jobs expire in seconds.
Because a Shortcut can do almost anything on a Mac or iPhone — control HomeKit, drive apps, run scripts, call web APIs — run_shortcut turns ChatGPT into a trigger for any automation you've already built.
For users — 3 steps
1. Create your account at shortcutsbridge.vercel.app and generate a pairing code.
2. Install the Mac agent (needs Node 18+):
npx shortcuts-agent pair <YOUR-CODE> # link this Mac to your account
npx shortcuts-agent install # keep it running on every restart
The first time a given shortcut runs, macOS may ask you to allow it — click OK (once per shortcut). It's per-shortcut consent, not a blanket automation grant.
3. Connect ChatGPT: Settings → Plugins → turn on Developer mode (Security & login) → Create a custom plugin → paste the MCP Server URL https://shortcutsbridge.vercel.app/mcp → Authentication OAuth → sign in & Allow.
That's it. In a new chat: "Search my Apple Shortcuts for anything about focus" or "Run my Morning Briefing shortcut."
Your Mac must be awake with the agent running.
npx shortcuts-agent installmakes it start automatically at login and restart if it ever crashes.
Managing the agent
npx shortcuts-agent status # is it paired & reachable?
npx shortcuts-agent logs # recent activity
npx shortcuts-agent uninstall # stop auto-start
What ChatGPT can do
| Tool | Action |
|---|---|
list_shortcuts | List every shortcut on your Mac (read-only) |
search_shortcuts | Find shortcuts by keyword (read-only) |
list_folders | List your Shortcuts folders (read-only) |
run_shortcut | Run a shortcut by name, with optional text input, and return its output |
The first three are read-only catalogue lookups. run_shortcut is the only tool that does something — it's open-world (the shortcut can take real-world actions), so ChatGPT confirms with you before running one, and only shortcuts you've already built can run.
Self-hosting
You can run your own relay so nothing depends on the hosted instance.
Prereqs: a Vercel account and a Supabase project (free tiers are fine).
- Storage — in your Supabase project's SQL editor, run
supabase/schema.sql. It creates a tiny Redis-shaped KV + queue surface (nb_*functions) with RLS on. - Deploy the
server/directory to Vercel. - Set env vars (see
server/.env.example):SUPABASE_URL,SUPABASE_SERVICE_ROLE_KEY— your project + service-role keyJWT_SECRET—openssl rand -hex 32
- Verify:
curl https://YOUR-APP.vercel.app/api/health→redisConfigured,redisOk,jwtSecretSetalltrue. - Point the agent at it:
npx shortcuts-agent pair <CODE> --server https://YOUR-APP.vercel.app.
How it works
- Auth — OAuth 2.1 with PKCE and Dynamic Client Registration (RFC 7591). ChatGPT registers itself, the user signs in on the ShortcutsBridge consent page, and ChatGPT gets a scoped MCP access token. JWTs are HMAC-signed; three kinds (
session,agent,mcp) with distinct privileges. Optional email verification via Resend (RESEND_API_KEY); off unless configured, and enforcement is opt-in (REQUIRE_EMAIL_VERIFICATION). - Relay — MCP tool call → job pushed to
jobs:<user>(TTL ≤ the caller's wait window, so an abandoned job can't run late) → the Mac agent (holding a hanging long-poll) is handed the job within ~200ms, executes it, pushes the result → the MCP handler returns it. The long-poll means jobs are delivered on arrival rather than on a fixed interval, so relay overhead is ~300ms instead of ~1s. The agent is considered "online" only while it's actively connected. - Agent — a dependency-free Node CLI. Tools shell out to the macOS
shortcutscommand-line tool (shortcuts list,shortcuts run); arguments are passed as argv (never shell-interpolated). A macOS LaunchAgent keeps it alive across logins and crashes.
Repository layout
server/ Vercel functions: OAuth + MCP endpoint + relay (deploy this)
agent/ shortcuts-agent — the npm-published Mac CLI (users npx this)
kit/ Browser automations: register the dev-mode connector AND
fill the OpenAI directory submission (see kit/README.md)
supabase/ schema.sql for self-hosting the storage
test/ end-to-end OAuth + MCP integration test
Development
# server
cd server && cp .env.example .env.local # fill in, then: vercel dev
# end-to-end test against a live deployment (reads ../.env.local)
node test/e2e-oauth.mjs
# agent unit tests
node --test agent/test-agent-unit.mjs
Security & privacy
- Shortcuts are listed and run only on your Mac. The relay temporarily stores requested content in short-lived job/result queues — job payloads live in Redis-shaped storage with a short TTL and are deleted on delivery.
- The agent token is stored at
~/.shortcutsbridge-agent.json(mode600). - No secrets are committed; see
.gitignoreand the.env.examplefiles. run_shortcutis marked open-world — a shortcut can take real-world actions — so ChatGPT asks before running one. Only shortcuts you've already built can run.
License
MIT © Isaiah Dupree
Production storage and owner verification
Relay 1.4 uses isolated service-role-only tables/functions on the configured shared Supabase project. Production refuses missing storage configuration, bounds RPC requests to ten seconds, and returns HTTP 503 when storage or signing configuration is unavailable. Apply the checked-in isolated-storage migration before deploying this version; it does not replace another bridge's account tables or functions.
node --env-file=.env.local test/live-storage.mjs verifies the actual persistent store, FIFO, expiry and counters using temporary integration keys. BRIDGE_SESSION_FILE=/private/owner-session.json BRIDGE_READ_TOOL=READ_TOOL node test/live-relay.mjs verifies the actual owner session, PKCE OAuth, hosted MCP and paired-Mac read; choose this bridge's inventory tool. No reviewer data counts as a live acceptance. Credentials and sessions belong outside Git.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.