WebMCP lets AI agents book tables, place direct orders, and search your menu from YOUR website. No commissions. No middlemen.
Run a 60-second audit to see your booking, menu, and visibility gaps before competitors in your neighborhood.
Free. 60 seconds. No signup required.
“Book me a table for 4 at a good Italian place near me, tomorrow at 7.”
Found: Your Restaurant — Direct booking via book_table
book_table( { date: "YYYY-MM-DD", time: "19:00", party: 4 })
Table for 4, Tomorrow 7pm — Direct booking. Zero platform fees.
Estimated annual commissions extracted from US restaurants by delivery and reservation platforms.
Estimate based on 15–30% take rates × ~$55B US food delivery market
The day Google launched WebMCP — a new web standard that lets AI agents book directly from restaurant websites.
Chrome 146 Early Preview · Backed by Microsoft
The commission you pay when a customer's AI agent books directly through YOUR site.
Google and Microsoft released WebMCP — a proposed web standard built into Chrome that lets AI agents interact with websites the way humans do: reading menus, checking availability, booking tables, placing orders. Except agents do it in seconds, with zero friction, and with structured precision.
Your restaurant's website can now publish a structured set of capabilities — not just for human eyes, but for AI agents. When a customer tells their AI “find me a good restaurant nearby,” the agent discovers YOUR tools, searches YOUR menu, and books directly through YOUR website. No OpenTable. No DoorDash. No commission.
This only works if YOUR website has WebMCP tools and your competitor's doesn't. The restaurants that implement first will be the ones AI agents find first. And once agents learn to use your tools reliably, they'll return — just like a regular customer who knows your name.
WebMCP is doing for AI agents what online menus did for Google Search. The restaurants that made their menus searchable early gained a decade of organic traffic. Structured AI tools may create a similar advantage.
Editorial observation, not a quote from any individual.
Each tool is a specific capability you give to AI agents. Some take 3 lines of HTML. Others need a bit of JavaScript. All of them mean more direct bookings and orders — without commissions.
Help customers find you — through their AI assistant, not through Yelp.
search_menuCustomer says
“What vegetarian options does your restaurant have?”
AI Agent
They have 8 vegetarian dishes, including Eggplant Parmigiana ($18), Wild Mushroom Risotto ($22)...
Why it matters: Your menu becomes searchable by ANY AI assistant. Customers with dietary needs find exactly what they want without scrolling your PDF menu.
get_hoursHTMLCustomer says
“Is your restaurant open on Sunday?”
AI Agent
Yes! Sunday hours are 11am–9pm. They’re currently open and have tables available.
Why it matters: No more lost customers calling a phone that nobody answers.
get_locationHTMLCustomer says
“Where is your restaurant? How far from downtown?”
AI Agent
your restaurant is at 123 Example St, 0.8 miles from downtown. Parking available on-site.
Why it matters: Works automatically if you already have LocalBusiness schema.org structured data on your site.
Direct bookings, zero commissions, full customer data.
check_availabilityCustomer says
“Do they have a table for 4 tomorrow at 7?”
AI Agent
Yes! They have availability at 7:00, 7:15, and 7:30 for a party of 4.
Why it matters: Your reservation system gets queried the same way it would if a human checked your website. But the agent does it in milliseconds.
book_tableHTMLCustomer says
“Reserve the 7pm table.”
AI Agent
I’m about to book: Table for 4 at your restaurant, tomorrow at 7pm. Confirm?
Why it matters: A new reservation appears in your system, tagged “Booked via AI Agent.” You have the customer’s info. What you don’t see: a $2.50 OpenTable commission.
modify_reservationCustomer says
“Can we change our reservation to 8pm?”
AI Agent
Done! Your reservation has been moved to 8pm. Confirmation sent.
Why it matters: No phone call needed. The agent handles modifications through your existing system.
Direct orders. Your food, your margin, your customer.
place_orderCustomer says
“Order me a margherita pizza, no olives, extra cheese, and a Caesar salad for delivery.”
AI Agent
Margherita pizza (no olives, extra cheese) + Caesar salad. Total: $28.50 + $3 delivery. Confirm?
Why it matters: You keep the full margin. A direct customer relationship. Their order history for remarketing. No 15–30% DoorDash commission.
track_deliveryCustomer says
“Where’s my food?”
AI Agent
Your order from your restaurant is being prepared. Estimated delivery: 35 minutes.
Why it matters: Every “where’s my order” handled by an agent is one fewer phone call to your kitchen.
Turn AI-driven customers into repeat customers.
check_rewardsCustomer says
“How many loyalty points do I have at your restaurant?”
AI Agent
You have 240 points — that’s enough for a free appetizer! Would you like to redeem on your next visit?
Why it matters: Loyalty programs become frictionless. Customers don’t need to remember their card or check an app.
get_specialsCustomer says
“What are the specials today?”
AI Agent
Today’s specials: Pan-Seared Branzino ($28) and Truffle Risotto ($24). Happy hour is 4–6pm with $8 cocktails.
Why it matters: Real-time specials push — agents always have current pricing and availability via provideContext().
10 tools shown · Your restaurant may need as few as 3 to get started
AI agents don't “prefer” your restaurant — they use whatever tools are available:
You have WebMCP. OpenTable doesn’t (yet).
AI agents book directly through your site. You pay nothing.
Both you AND OpenTable have WebMCP.
Agents compare both. Your tool has your story, your atmosphere. Better tools win.
OpenTable has WebMCP. You don’t.
AI agents book through OpenTable. You pay commissions. Again.
The window for Scenario A is now.
OpenTable, Resy, and Yelp will eventually add WebMCP. Restaurants that implement first establish the AI booking channel before the intermediaries reclaim it.
The customer’s AI agent discovers your restaurant has a ‘check_availability’ tool on your website.
The agent calls your tool with the date, time, and party size. Your system responds with available slots — in milliseconds.
The agent fills your reservation form. WebMCP requires customer confirmation before submitting — the customer sees a summary and taps Confirm.
The booking enters your system. You have the customer’s name, phone, and party details. No commission. Next time they ask, your restaurant is in the agent’s memory.
The US food delivery market is approximately $55 billion annually. Third-party platforms typically take 15–30% of every order. For a restaurant doing $20,000/month in delivery, that can mean $3,000–$6,000/month in estimated commissions. WebMCP changes the equation.
AI Agents Don’t Just Read Your Menu — They Understand It
Your menu becomes structured data that agents can search, filter, and reason about. When a customer says “I want something spicy and gluten-free under $20,” the agent queries your menu tool with structured parameters:
Agent sends:
{ dietary: "gluten-free", flavor: "spicy", max_price: 20 }Your tool returns:
[
{ name: "Thai Basil Chicken", price: 16,
tags: ["gluten-free", "spicy"] },
{ name: "Green Curry Shrimp", price: 18,
tags: ["gluten-free", "spicy"] }
]Key insight: PDF menus are invisible to AI. Structured menus are searchable, filterable, and comparable. The restaurant with a structured menu gets the order.
Hypothetical Example
Actual savings depend on your delivery volume, commission rates, and AI agent adoption in your market.
WebMCP sits between your website and your existing systems. It doesn't replace Toast or Square — it gives AI agents a structured way to interact with what you already use.
Customer's AI Agent
“Book me a table for 4 at 7pm”
WebMCP Plugin / Code Layer
Reservation System
OpenTable / Resy / In-house
Ordering System
Toast / Square / Custom
The AI agent talks to your website. Your website talks to your existing systems. A reservation from an AI agent looks the same as a reservation from your website — because it is one.
When an AI agent places an order through your website, the order flows into your POS exactly like any other online order. Zero retraining.
The booking still comes through YOUR website (0% commission). The integration just syncs with your calendar so you don’t double-book.
✔ Available now: CMS plugins (WordPress, Squarespace, Wix, Shopify), custom site SDK, Declarative HTML tools, Imperative JS tools, Agent Readiness Score, basic analytics.
● Coming Q2 2026: Toast POS integration, Square POS integration, reservation system sync, delivery API integrations.
● On roadmap: Clover, Lightspeed, additional POS systems, advanced analytics, multi-location management.
Even without POS integration, restaurants can start today: your existing website forms + WebMCP Declarative attributes = agent-ready in minutes.
WebMCP launched two weeks ago. We don't have case studies yet. Here are three estimated projections based on typical restaurant industry economics. All numbers are hypothetical.
Hypothetical: independent restaurant, 1 location
Assumptions
Estimated Month 1 (5%)
Estimated Month 6 (15%)
Based on these assumptions, estimated annual savings range from $1,800 to $5,400. Actual results depend on AI agent adoption rates in your market.
Hypothetical: fast-casual, heavy delivery volume
Assumptions
Estimated Month 1 (5%)
Estimated Month 6 (20%)
These are projections based on assumed adoption rates. AI agent ordering is new and actual adoption will vary by market and customer behavior.
Hypothetical: 15 locations, dine-in + delivery mix
Assumptions
Estimated Month 1 (5%)
Estimated Month 6 (15%)
Multi-location groups have more to gain from commission reduction, but these estimates assume adoption rates that are not yet proven.
These are projections, not results.
AI agent adoption is in its earliest stages. We don't know how quickly consumers will use AI agents for restaurant bookings. What we do know:
Be among the first 100 restaurants to go agent-ready.
If you can install a WordPress plugin or add an app to your Shopify store, you can make your restaurant agent-ready.
2 minutes
Install the WebMCP plugin from your CMS app store. WordPress, Shopify, Squarespace, Wix — we support them all. One click.
3 minutes, automatic
Our AI Co-Pilot crawls your website automatically. It finds your menu, your reservation form, your hours, your contact info.
3 minutes
Review tool descriptions in plain English. Toggle tools on or off. Add your restaurant’s personality to descriptions.
2 minutes
Hit Activate. Your restaurant is now discoverable by AI agents. Your Agent Readiness Score appears instantly.
Your score appears here
Compare these prices to what you're already paying platforms.
Menu browsing + hours + basic presence
Enough to start appearing in AI agent results.
Direct AI bookings, zero commissions
vs. per-cover fees on reservation platforms
Designed for restaurants looking to reduce per-cover platform fees.
Ordering + loyalty + full analytics
vs. percentage-based delivery commissions
Designed for restaurants with significant delivery volume.
Multi-location, POS, enterprise
We built WebMCP for the restaurant that doesn't have a CTO. For the independent operator who wants to be found by AI without hiring a development team.
Getting found by AI agents shouldn't require a major technology investment. It should take a few minutes and a plugin. That's what we're building.
When a customer asks their AI “What's good at that Italian place?” — the agent needs data, not a photo of a laminated card.
Margherita Pizza.......$16
Pepperoni Pizza........$18
Wild Mushroom Risotto..$22
Caesar Salad...........$12
RESULT: UNREADABLE
{
"name": "Wild Mushroom Risotto",
"price": 22.00,
"dietary": {
"vegetarian": true,
"gluten_free": true
},
"allergens": ["dairy"],
"available": true
}RESULT: FULLY STRUCTURED
Restaurants aren't static. Items run out. Specials change daily. WebMCP handles this through provideContext() — your website tells AI agents what's current, right now.
Chef updates daily specials
Agent: “Today’s specials are Pan-Seared Branzino ($28) and Truffle Risotto ($24).”
Happy hour starts
Agent: “Happy hour is on now! $8 cocktails and $6 draft beers until 6pm.”
Kitchen 86’d the branzino
Agent: “The branzino is sold out tonight, but they have Grilled Swordfish and Salmon Piccata.”
Peak hour — 45 min wait
Agent: “Current wait time is about 45 minutes. Shall I book for later tonight?”
For the estimated 32 million Americans with food allergies (per FARE), finding safe restaurant options through AI agents is not just a convenience — it can be a safety feature.
“I need a nut-free restaurant for my daughter”
Agent filters entire menu by allergens and confirms safe options exist before recommending.
“What can I eat here if I’m vegan?”
Agent returns all vegan items, including possible modifications: “The Caesar can be made vegan.”
“I’m celiac — is the pasta gluten-free?”
Agent checks: “They offer gluten-free penne as a substitute. The risotto is naturally GF.”
Potential advantage: Restaurants with richer dietary data may become preferred recommendations for customers with dietary needs. If your competitor's menu just says “vegetarian: yes/no” and yours includes allergens, modifications, and preparation notes — the agent has more reason to recommend you.
You don't need to create structured data yourself. Here's how it works:
The JSON, the schema, the provideContext() calls — all of that happens under the hood. You never see it.
Restaurants that prepare for AI agent interactions early will be better positioned when agent-assisted dining decisions become common.
Enter your URL to get a concrete scorecard for menu structure, booking flow, hours visibility, and schema gaps in under 60 seconds.
See Your Agent ScoreThen activate the plugin and publish your first live tools in about 10 minutes.
For multi-location brands, we audit each site, map POS/reservation stack differences, and deliver a rollout plan by location.
Talk to Our TeamIncludes timeline, integration priorities, and expected savings range per market.
Ship agent-readiness as a service with production templates for tools, validation, and reservation/order safety patterns.
View TemplateStart with the template, then move to white-label workflows for your client portfolio.
Feb 9, 2026
WebMCP launches in Chrome 146
◀ You are here
Q2 2026
Early adopters see first results
Q3–Q4 2026
Platforms catch up (OpenTable, Resy add WebMCP)
The window between “nobody has this” and “everybody has this” is measured in months.
Every platform shift in the restaurant industry has created opportunities. Restaurants that adopted online menus early gained organic traffic. Those that built strong Instagram presence built lasting brands. Those that partnered early with delivery platforms got volume — but often at the cost of their margins.
This time is different. WebMCP doesn't ask you to give away your margin to a platform. It gives your customers' AI agents a direct line to YOUR kitchen, YOUR host stand, YOUR menu.
The question is straightforward: when customers start asking AI where to eat, will your restaurant be ready to answer?
Free. 60 seconds. No signup required.