🍜

Ramen Ichiban
24/7 Voice Line

Every small business deserves a voice assistant
that never sleeps, speaks every language, and never gets an order wrong.

DeepLearning.AI Builder Night Tokyo

The Problem

📞

Missed calls = missed revenue

During lunch rush, after hours, or when the staff is busy — the phone rings and nobody picks up. Every missed call is a lost customer.

🌍

Language barriers in Tokyo

40M+ tourists visit Japan yearly. They can't read the menu, can't explain allergies, and can't make reservations — all because of language.

⚠️

Allergen communication failures

Miscommunicated allergies are dangerous. Pointing at a menu and hoping the staff understands "no peanuts" isn't good enough.

📝

Manual reservation chaos

Paper reservation books, missed bookings, double-bookings. Small restaurants don't have systems — they have sticky notes.

The Solution

One voice agent that
finds, informs, and books
in any language, 24/7

📍

Find Nearest Outlet

"I'm near Shibuya Station" → gets the 3 closest locations with hours

🥢

Menu & Allergens

"I'm allergic to pork" → instantly knows what's safe and what's not

📅

Live Reservations

Checks real availability, books on Google Calendar, gives confirmation

How It Works

Customer (Voice)

↕

Vocal Bridge Agent

↕ MCP Protocol

MCP Server + Worker

↕

Reservations Database Google Maps API Google Calendar

5 MCP tools · Google APIs for location & booking

The MCP Server

One server. Five tools. The voice agent calls them as needed — no hardcoded flows.

📍 find_nearest_outlets

Takes a location name or coordinates. Geocodes via Google Maps Places API, queries the database, calculates distance, returns the 3 closest outlets.

in: "Shibuya Station" out: Roppongi 0.3km, Shibuya 1.2km...

🥢 get_menu

Returns the full menu for any outlet — items, descriptions, prices, grouped by category. The agent summarizes instead of reading the whole list.

in: outlet_id out: 18 items across 4 categories

⚠️ check_allergens

Cross-references a customer's allergies against every menu item. Returns two lists: safe to eat and unsafe with the specific allergen that matched.

in: ["pork", "dairy"] out: 8 safe, 7 unsafe with reasons

🕐 check_availability

Reads the outlet's Google Calendar, compares against business hours, and returns open 1-hour slots for the requested date and party size.

in: date, party_size out: 7 available slots

📅 make_reservation

Writes to the database first (guaranteed), then syncs to Google Calendar. Returns a confirmation ID. The dashboard shows it in real time.

in: date, time, name, party_size out: Confirmation #A3F8B2C1

Live Demo

1

Find
Locate nearest outlet from any location

2

Explore
Browse menu, get recommendations

3

Check
Verify allergens, find safe options

4

Book
Reserve a table, get confirmation

Tech Stack

🎙️

Vocal BridgeVoice agent platform — Chatty mode, MCP integration, multilingual

⚡

Cloudflare WorkersMCP server — serverless, edge-deployed, global latency

🗄️

Cloudflare D1SQLite at the edge — outlets, menus, allergens

📍

Google Maps Places APINew API — geocoding user locations to find nearest outlet

📅

Google Calendar APIReal-time availability checks and reservation booking

🌍

30+ LanguagesAuto-detects and responds in the customer's language — Japanese, English, and more

Let's Talk Numbers

40M+

tourists visit Japan yearly
who can't read a menu

44m

built in 44 minutes
from zero to live demo

30+

languages supported
out of the box

Built in 44 minutes with Vocal Bridge + Cloudflare + Google and Vibes.
Today it's a ramen shop. Tomorrow it's your dentist, your hotel, your salon.
Same stack. Different prompt. Every missed call is someone else's customer.