Introduction

Vio is a hotel metasearch platform that aggregates offers from multiple providers — including Booking.com, Expedia, Hotels.com, and Agoda. The Vio MCP server exposes this data layer to AI agents through the Model Context Protocol (MCP).

Why hotel search in your AI agent?

Travel is a natural vertical for AI assistants. Users ask conversational questions — "pet-friendly hotel in Barcelona with a pool, under 150 per night" — that map directly to structured hotel searches. Adding hotel search to your agent means:

• High engagement — Travel queries are action-oriented. Users searching for hotels are ready to compare, decide, and book.
• Natural fit for AI — Hotel search involves multi-criteria filtering, date flexibility, price comparison, and personalization — tasks where AI agents excel over traditional search UIs.
• Monetization — Hotel bookings generate revenue through commission-based attribution. Every booking your agent drives is tracked and compensated.
• Retention — Users who complete high-value tasks through your product come back. Hotel search creates a reason to return.
• Accuracy — LLM knowledge and web search cannot reliably provide real-time hotel availability, current pricing, or bookable offers. Hotel data changes constantly — rates shift by the minute, rooms sell out, and promotions expire. A structured data source is the only way to give users results they can act on.

Why Vio?

Vio handles the hard part — aggregating rates from multiple providers in real time, normalizing data across sources, and serving it through an API designed for AI agents.

• Best available rates — Real-time price comparison across all major booking platforms
• Comprehensive data — Details, reviews, AI-generated insights, availability calendars, price analytics
• AI-optimized — Semantic search, self-correcting errors, and progressive data loading designed for LLMs
• Pre-built UI — Drop-in hotel cards, maps, and booking flows via @vio/ui-kit — or embed a fully hosted app with zero code
• Zero maintenance — Tools are discovered via MCP; schema changes are picked up automatically
• Revenue sharing — Booking attribution and commission for every partner integration

How it works

Your AI agent connects to the Vio MCP server, discovers the available tools, and calls them as needed during conversations. Vio aggregates offers from multiple providers and returns structured results — your agent handles the rest.

1. Your AI discovers available tools and their schemas via tools/list
2. A user expresses intent in natural language (e.g., "pet-friendly hotel in Barcelona under 150/night")
3. Your AI translates that intent into structured tool inputs
4. The Vio MCP server queries multiple providers and returns structured results
5. Your AI presents results to the user — as text, cards, or custom UI

Integration options

Choose the level of depth that suits your product:

Integration	What you get	Time to integrate
MCP Only	Data layer — you build all UI	~1 day
UI Kit Library	Data + pre-built React components	~2-4 hours
AG-UI + UI Kit	Data + UI + automatic interaction handling	~2 hours
Hosted MCP App	Everything — embed an iframe, done	~30 minutes

All modes call the same MCP tools. Start wherever makes sense and migrate upward without reworking your data integration. See Integration Options for details.

Live example

The Vio MCP server powers the Vio ChatGPT App — a live integration with interactive hotel cards, maps, and booking flows.

What Vio provides

• API key and endpoint access
• Documentation and integration support
• Pre-built UI components (@vio/ui-kit) or a fully hosted MCP App
• Revenue sharing on bookings
• Analytics dashboard for monitoring usage and conversions
• Ongoing support and feature development

Contact your Vio representative to discuss API key provisioning, integration approach, and revenue sharing terms.

Next steps

• Features — See how it fits into your product
• Quick Start — Make your first API call
• Integration Options — Choose the right integration mode
• API Reference — Detailed tool documentation