An open-source project called Travel Hacking Toolkit gives AI coding agents like Claude Code and OpenCode the ability to search award flights across 25+ loyalty programs and automatically figure out whether you should burn points or pay cash. Built by developer @bartholomej, it combines MCP servers (which give AI agents real-time tool access) with markdown-based "skills" that teach Claude Code and OpenCode the ability to search award flights across 25+ loyalty programs and automatically figure out whether you should burn points or pay cash. Five servers work immediately with no API keys: Skiplagged, Kiwi.com, Trivago, Ferryhopper, and Airbnb. Premium integrations go further. Seats.aero ($8/month) handles award availability, and SerpAPI compares cash prices.
The core workflow is straightforward. Ask your AI agent to find a 60,000-mile business class flight to Tokyo. It searches award availability, pulls cash prices from Google Flights, checks your loyalty balances via AwardWallet, and recommends the best booking strategy. The project includes data files with airline alliances, hotel chains, transfer partners, and "sweet spot" award redemptions that agents reference when making decisions. Users will need to add their own API keys for the premium services.
Hacker News commenters pointed out some real limitations. Families struggle with award availability since finding 4+ seats on the same flight is tough. Others noted the toolkit could benefit from integrating credit card portal benefits like Amex Fine Hotels & Resorts, which throw in credits and perks that change the math on points versus cash. The project also faces ongoing maintenance challenges since travel APIs change frequently and without warning.