FLIGHT DATA / MCP
Your agent asks.
The source matters.
Give an AI travel product a structured way to request airline fare observations. Farelint’s MCP server uses the same workspace, source capabilities and credit rules as the HTTP API.
How the integration works
- Check capabilities and balance. Use
farelint_list_airlinesandfarelint_get_usage. - Ask for the missing trip details. Confirm airports, dates, passengers and a supported source before collecting data.
- Submit once, then poll.
farelint_search_flightscreates a job. Retrieve its outcome withfarelint_get_job. - Explain the result honestly. Retain capture time, currency, price basis, unknown fields and checkout-verification status.
One endpoint. Your existing API key.
https://api.farelint.com/mcp
Authorization: Bearer YOUR_FARELINT_API_KEYThe server uses Streamable HTTP with bearer API-key authentication. The integration guide includes Claude Code and Cursor configuration. Clients must support supplying a bearer key; OAuth sign-in is not implemented. Never put a secret in a public prompt, web page or client-side application.
What can an agent do?
Inspect airline capabilities and usage, submit a supported flight search, poll a job and inspect workspace request history. Offer-details requests are available only when enabled for the selected source. The agent does not get wider coverage or extra permissions through MCP.
A tool call is not a booking. A queued job is not a fare quote. A completed observation is not a checkout guarantee. Collection requires an activated workspace and available workers.
Build recommendations with visible evidence
For a planning assistant, the useful part is the record behind its answer: which trip was checked, when the price was captured, what it covers and what is unknown. Inspect the example data before integrating.
MCP is an interface, not an exclusive feature or a promise of instant results. Packaged skills, scheduled data feeds and historical analytics remain separate roadmap items.