A restaurant manager built a booking system that any AI assistant can search for live tables, and it deliberately never books. Every salon, clinic and bar that takes bookings will need this front door
A Sydney restaurant manager who taught himself to code built a booking system that any AI assistant can search for live tables and appointments, and on purpose it never books anything. Most venues are invisible to assistants today, and making them findable is a business.
Published Nov 5, 2026
The Wow
What someone built
Gaplessly is a booking system for salons, clinics, restaurants and bars. Its public MCP server is one endpoint with no API key or sign-in: an assistant can find a venue, read its local clock and list open table or appointment times. We listed the live server's tools on Nov 5, 2026.
- Original source
- Gaplessly MCP server docs
- License
- Hosted service; link only. The CLI is published on npm and GitHub (github.com/gaplessly/cli); check its licence before reuse
- What you need
- An MCP-capable assistant
- A venue booking system with an API (to build your own)
- Phone testing of the booking link
- Install
- claude mcp add --transport http gaplessly https://gaplessly.com/api/mcp (no key, no sign-in), then ask for a table at the demo venue Saltline.
- A business you could build on it
- Assistant-ready booking setup for local venues
Why it's remarkable
The last step is a link, not a booking. No assistant can prove who it is acting for, so no tool takes a name, email or phone; the guest taps a short-lived link to the venue's own booking page, which holds the table for about ten minutes. Each time comes with a signed slot token that expires in about 30 minutes and fails if a prompt-injected request tries to move it to another venue.
The Opportunity
The business it reveals
Every venue that takes bookings will want assistants to find its open times, and few owners can build it. A one-person studio can make venues in one vertical assistant-ready with the same find-then-hand-off pattern, starting with a free check of what ChatGPT or Claude can see today.
Who would pay
Independent restaurants, salons, clinics and studios that already take online bookings; later, small vertical booking-software vendors.
The Flip
2 ways you could turn this into a business
Assistant-ready booking setup for local venues
- Who pays
- Independent restaurants, salons, clinics and studios
- How it would earn
- A setup fee to make the venue's live times findable by AI assistants with a safe handoff, then a small monthly fee to keep it working (our hypothesis)
- Why it might work
- The pattern is documented and running, the endpoints are small, and most venues have no idea whether assistants can find them.
- What already exists
- Waiting for the venue's booking platform or Reserve with Google to add it; marketplaces that take a commission per booking.
A find-then-hand-off MCP server for one niche booking platform
- Who pays
- Small vertical booking-software vendors (yoga studios, tour operators, tradespeople)
- How it would earn
- A build fee or licence for an MCP server on their API that follows the same safe design (our hypothesis)
- Why it might work
- Vendors want assistants to send them customers but few have built a server, and Gaplessly published the design decisions that matter.
- What already exists
- The vendor building it in-house, generic API-to-MCP wrappers.
Why we'd try "Assistant-ready booking setup for local venues" first: Every venue that takes bookings will want to be findable by assistants, owners can't build it themselves, and a free test of 'can ChatGPT find your open tables?' is an easy first conversation.
Build It
How you could build it
A read-only MCP server (or structured booking pages) for one venue or one booking platform that returns live times with signed, short-lived slot tokens and hands the guest a link to the venue's own booking page.
First step: Add Gaplessly's MCP server to Claude or Cursor and ask for a table for two tomorrow at Saltline. Watch what comes back: display-only times, a signed token, then a link. Then pick five venues near you and ask your assistant to find each one's next open time.
Agents and tools
- Gaplessly's server and docs as the working reference
- Claude Code, Codex or Cursor
- The venue's booking system API
- A small hosting tier for the server
Setup
- Read Gaplessly's build log and docs, and try its five tools on the demo venue.
- Check the venue's booking system for an availability API and its terms.
- Build find-times and start-booking tools that return display-only times and signed tokens.
- Make the last step a link to the venue's own booking page.
- Test from two assistants and a phone before going live.
What the agent does
- Writes the server code and tool descriptions
- Answers guests' availability questions
- Drafts the venue's assistant-facing docs
What you do
- Sells to venues and checks platform terms
- Tests the handoff on real phones
- Supports the venue when something breaks
Your first customer: A venue you already visit that takes online bookings, offered a free check of whether AI assistants can find its open times.
Test demand first: Run the prompt on ten venues in one suburb over a week and show each owner what an assistant sees today. Success: three reply and one asks what setup would cost.
Your starter instructions
Paste this into the AI agent you use (ChatGPT, Claude, Gemini, Grok Bot or another) as your first message. Copying it starts work in your own agent; it doesn't create or deploy a business by itself.
Running it
What stays human: Picking the vertical, selling to venue owners, checking each booking system's terms, testing the handoff on a phone, and supporting the venue when an assistant gets something wrong.
- Startup cost (estimate)
- Under $25 to start (estimate) (Estimate: the Gaplessly endpoint is free to try with no key (checked Nov 5, 2026); a small MCP server can run on a free or low hobby tier. The rest is an AI plan you may already pay for)
- Running cost
- Hosting in single dollars a month per venue plus your time (estimate) (Estimate: a read-only availability server makes a few calls per conversation; hours unmeasured)
- How much the agent runs
- Agents run one core part; people run the rest.
Pricing your own version? Use the free startup-cost worksheet.
The evidence
What the source shows: Gaplessly's DEV build log (Nov 3, 2026), its MCP docs and its live endpoint, checked Nov 5, 2026. Gaplessly is in early access and publishes no customer or revenue numbers.
What's ours: The setup service, its buyers and the flips are our application of the pattern.
Worth planning around
- Live times depend on the venue's Gaplessly plan; other venues return a booking-page link.
- Assistant-driven bookings are still a small share of demand, so the first buyers are owners who want to be early.
- Each booking platform's API and terms decide what a third party can build.
Sources
- Gaplessly on DEV: "I built an MCP server for restaurant bookings, and it never books anything" (Nov 3, 2026)
- Gaplessly MCP server docs (checked Nov 5, 2026)
Sources reopened Nov 5, 2026. Spot a mistake? Tell us and we'll correct it in public.
Keep reading
- How people make money with n8n (guide)
- Cursor: plans and what people build and sell with it (guide)
- How much of the work do AI agents actually run? (guide)
- He still can't write code. His five-person ad agency now runs on about 270 tools he built with Claude. Every owner-run firm needs its first five
- A marketer who doesn't code built an elk-call coaching app with Claude Code and says it hit a $650K run rate in 110 days. Every skill learned by sound is the same opening
- A free n8n workflow flags fake supplier invoices before anyone pays them — and Germany's e-invoice switch makes it a service bookkeepers can sell
- Marc Lou's agent chased a broken-shoe refund through three support desks and got his 850 AED back. Small stores now need a claims desk that can answer agents
- Every discovery and launch plan