Ask Claude About Your Venue.
It Can Actually Answer.
Connect your Everybooking account to Claude, Cursor, or any MCP client, and the assistant stops guessing. It reads your live bookings, your real pricing, your actual availability. 400+ tools over the same data and permission model your staff use. And every edit it makes is version-tracked and revertible.
Already have an account? Setup guide →
An assistant that can look things up.
You ask an assistant about your venue and it answers from nothing. It has never seen your calendar, your packages, or the inquiry that came in this morning.
- You paste in a spreadsheet, then paste the answer back
- Every answer is a guess dressed as a fact
- Nothing it suggests can be applied without you doing it
- The context dies when you close the tab
The assistant queries your account directly, gets the real answer, and can make the change you asked for while you watch.
- Reads live bookings, catalog, pricing and availability
- Writes back: quotes, products, funnel steps, pricing
- Every edit versioned, with undo and conflict detection
- Scoped to your account by the token you mint
- Nothing installed — a remote endpoint, not a local process
Things venue operators actually ask.
“Which weekends in October still have the barn open?”
Reads live availability across your real inventory, not a synced copy.
bookings“Raise every bar package by 8% for next season, but leave the non-alcoholic one alone.”
Edits pricing across the catalog in one pass, instead of one product at a time.
catalog“Who inquired about a December wedding and never got a quote?”
Filters the pipeline on the question you actually have, rather than the reports someone built in advance.
crm“Draft a quote for the Moreau group — 48 guests, two nights, full board.”
Builds the multi-line quote against your rules and drops it into the deal for you to review.
bookings“Add a Saturday-only brunch step to the retreat funnel.”
Authors the funnel step directly. Every change is versioned and revertible.
funnels“What did we book last October, and how does this October compare?”
Answers across the whole account without exporting anything to a spreadsheet.
operationsEvery AI edit is a reversible edit.
The reasonable objection to letting an assistant touch a live booking system is: what happens when it gets something wrong.
Everybooking treats an MCP edit exactly like a human one. It is recorded in the same version history, it shows in the same audit trail, and it comes back with the same undo. Human edits, in-app AI edits, and MCP edits are all tracked together, with conflict detection when two of them touch the same thing. An assistant working on a quote writes into the same draft you are typing in, and it appears in your open composer in about a second — so you are watching it work, not discovering it afterwards.
If you want a harder boundary than undo, narrow the surface. Point the client at a toolset and everything outside it is invisible to the assistant — a read-only research setup and a full operations setup are the same endpoint with a different parameter.
Three steps, one command.
Mint a token
Settings → API Tokens in your Everybooking account. The token is what selects the account, so there is no separate endpoint to find.
Point your client at it
One shared host for every account: https://mcp.everybooking.com/mcp, authenticated with a bearer token.
Ask it something
Start with a question you already know the answer to. It is the fastest way to confirm it is reading the right account.
For Claude Code, the whole setup is one line:
claude mcp add --transport http everybooking \
https://mcp.everybooking.com/mcp \
--header "Authorization: Bearer YOUR_TOKEN"Claude desktop, Cursor and other clients take the equivalent entry with a URL and an Authorization header. Full setup and troubleshooting →
Give it as much as you want it to have.
Add ?toolset=bookings,catalog to the endpoint and the assistant sees only those tools. Useful for keeping a research assistant away from anything that writes.
bookingsDeals, quotes, availability, depositscatalogProducts, packages, pricing rules, inventorycrmContacts, companies, pipeline, historyfunnelsBooking funnels and their stepsoperationsDay-to-day running of the accountvoiceCall handling and transcriptsdiagnosticsHealth and troubleshootingfullEverything, the defaultIf your bookings are single-slot, one price, one person, this is more machinery than the job needs, and a calendar tool will serve you better. The venues that get value here are the ones where a booking is a real quote: ten to a hundred line items, a group to coordinate, deposits to split, and enough moving parts that answering “what did we charge the last group like this one?” currently means opening four tabs.
MCP access needs a Growth plan or higher. See pricing →
MCP, answered.
- What is MCP?
- MCP (Model Context Protocol) is the open standard for connecting an AI assistant to a real system. Without it, an assistant can only talk about your venue. With it, the assistant can actually read your bookings and change them, using the same permissions your own staff have. Everybooking runs a remote MCP server, so there is nothing to install and nothing running on your machine.
- Is this different from the AI already inside Everybooking?
- Yes, and they are for different jobs. The quote agent inside Everybooking runs on its own, answering inquiries the moment they land whether or not anyone is at a keyboard. MCP is for when you want to sit down and work with an assistant directly, in your own words, across the whole account at once. The quote agent handles the inbound. MCP handles the questions and the bulk work you would otherwise do by clicking.
- What if the AI gets something wrong?
- Every edit made through MCP is version-tracked the same way a human edit is, with undo and conflict detection. You can see what changed, who or what changed it, and revert it. This is the part most AI integrations do not have, and it is the reason it is safe to let an assistant touch a live booking calendar rather than a copy of it.
- Which AI assistants work with it?
- Any MCP client. Claude (desktop, web and Claude Code) and Cursor are the ones venue operators use most, but the endpoint is a standard Streamable-HTTP MCP server and is not tied to one vendor. You authenticate with an API token from your own account, so the assistant only ever sees your data.
- Do I need a developer to set this up?
- For Claude Code it is one command. For Claude desktop or Cursor it is pasting a URL and a token into a settings panel. Minting the token takes about a minute in Settings → API Tokens. The full walkthrough, including troubleshooting, is at /docs/mcp.
- Which plan do I need?
- MCP access requires a Growth plan or higher. Free and Starter accounts get a 402 from the endpoint. API and MCP write access is also included on Pro and Enterprise.
- Is my data safe if an assistant can reach it?
- The token you mint scopes to your account and nothing else — the server enforces tenant isolation, so one account's token can never read another's data, and browser-based clients are restricted by an Origin allowlist. You can also narrow what the assistant can touch at all: pass a toolset parameter and the assistant only ever sees, say, bookings and catalog, with the rest of the surface invisible to it. Revoke the token and access ends immediately.
See what it would run on.
Enter your website and watch a booking funnel build from your own services. That funnel, your catalog and your pipeline are what the assistant reads once MCP is connected.