How We Built an MCP Server for Private Aviation (Claude Integration)
When Anthropic released the Model Context Protocol (MCP), we saw an opportunity to make VOLO the first private aviation platform natively accessible from Claude Desktop. Here's how we built it.
Why MCP Matters for Aviation
MCP allows AI assistants like Claude to use external tools as naturally as they use their own capabilities. For private aviation, this means a user can simply ask Claude: "Find me a jet from Tokyo to Singapore next Thursday for 6 people" and Claude can directly search our fleet, compare options, and initiate a booking.
This is fundamentally different from a REST API. With REST, a developer needs to write code to integrate. With MCP, the AI assistant IS the integration layer. No code required from the end user.
Tool Design
We designed five MCP tools for the VOLO server:
- search_flights: Natural language flight search. Accepts queries like "heavy jet, NYC to London, March 20, 8 passengers" and returns structured results with aircraft details, pricing, and availability.
- get_quote: Generates a detailed, all-inclusive quote for a specific aircraft and route combination.
- check_availability: Real-time availability check for a specific aircraft type on given dates.
- list_aircraft: Browse available aircraft by category with full specifications.
- get_empty_legs: Find discounted empty leg flights on specific routes or date ranges.
Natural Language Processing
The key engineering challenge was making the tools work with the messy, ambiguous input that real users provide through Claude. "I need to be in Aspen by Friday" is very different from a structured API call, but our NLP layer handles the translation.
We process date references ("next Friday", "in two weeks", "March 20th"), location ambiguity ("Aspen" vs "ASE" vs "Pitkin County Airport"), and implicit requirements ("family trip" implies needing more cabin space).
Architecture
The MCP server runs as a lightweight Node.js process that Claude Desktop spawns locally. It communicates with VOLO's backend API over HTTPS, handling authentication, rate limiting, and response formatting.
We chose to keep the server stateless — each tool call is independent. This simplifies deployment and means the server can be updated without losing conversation context in Claude.
Results
Since launching the MCP server, we've seen a 3x increase in API engagement from Claude-based agents compared to our REST API alone. The natural language interface dramatically lowers the barrier to accessing private aviation data.
If you're building an AI agent that handles travel or luxury services, our MCP server is the fastest way to add private aviation capabilities. Install it from our GitHub repository and your agent can search flights within minutes.