Last updated: 2026-06-01
Trip Quill uses AI to extract itinerary details from travel documents you upload (PDFs, screenshots) and to summarize destinations. We process this data through external AI providers; this page tells you exactly which ones and what they receive.
Read Google's data-handling policy for the Gemini API: https://ai.google.dev/gemini-api/terms
When you upload a booking document, the bytes are sent to Google's Gemini API along with a short instruction prompt. Google's standard policy for the Gemini API applies (Google does not use your inputs to train their models under that policy as of the date above).
When the backend synthesizes a trip view, the per-asset structured extracts (text only — no original files) are sent to Gemini, along with your display name and home country so the AI can identify you as a passenger in your own documents and write the summary in your language. On a shared trip, the summary is written once for everyone on it: the members' first names (full names when two share a first name) are sent instead, so the AI can tell the travellers apart, and the summary is written in the trip creator's language.
For destination curator notes, only the destination code and a small set of public facts (capital, currency, driving side, official languages) are sent. No user-identifying data.
Your account, trip metadata, and all original asset bytes you upload live in Trip Quill's own infrastructure (Cloudflare R2 for files, Postgres for metadata). Cached AI outputs (extracts, synthesis, curator notes) live in Trip Quill's Postgres too.
If we add or change a provider, this page updates and your consent is re-prompted on the next sensitive action (asset upload). The backend will refuse to start a release that adds a provider without first updating this disclosure.
Questions? Email privacy@tripquill.app.