12 honest answers

Yes — every itinerary, destination guide, and specialist vertical page (accessibility, dietary, visa) is free to read. No paywall, no “free tier with asterisks”.

We plan to monetize in three ways: display ads via Google AdSense (loaded only after you accept cookies), affiliate commissions on hotel / flight / tour bookings you choose to make through our links (at no extra cost to you — see our affiliate disclosure), and an optional premium tier with unlimited personalized itineraries and PDF export. The reference content — destinations, verticals, all 172 city guides — stays free forever.

Honest answer: better than nothing, but not a substitute for checking primary sources close to your trip.

We use Google Gemini Flash Lite Preview grounded in a curated destinations database — 172 cities with real GPS coordinates, live weather data, average trip length, and “popular for” interest tags. We add Book ahead warnings on attractions known to sell out 1-3 months in advance (teamLab Planets, Louvre, Colosseum).

What LLMs still get wrong: specific opening hours, small restaurant details, seasonal closures. We prefer generous ranges (“most mornings 10-18”) over exact times and we actively tell you to verify before booking. When we’re uncertain, we say so in the copy instead of faking confidence.

Because in these categories, a wrong answer causes real harm: missed flights because your passport didn’t need the ESTA you skipped, a wheelchair user arriving at a “fully accessible” restaurant that has four steps, or a coeliac ordering a dish AI called gluten-free when local restaurants cross-contaminate routinely.

So we run these as firewall verticals:

• Visa matrix: 81 curated entries across 9 passports and 8 destinations, every one linked to an official government source (Japan MOFA, Thailand MFA, UAE MOFAIC, Singapore ICA, etc.) with a visible last-verified date. Quarterly manual review. No AI generation at all.
• Accessibility guides: OSM wheelchair=* tags (real humans tagging venues) plus Wikipedia context + a named human editor who rates neighborhood terrain honestly.
• Dietary guides: OSM diet:halal|kosher|vegan|gluten_free tags plus local-dish compatibility ratings (“always” / “usually” / “ask” / “never”).

AI is used where the structure is repetitive and the stakes are low (packing lists, itinerary sequencing, seasonal recommendations). Humans handle the high-stakes layers.

Tripozi is a solo-founder indie project built in April 2026 in the spirit of asking “what does an honest travel reference look like if we stop pretending it needs to cover every long-tail keyword?”. The operator is a registered Polish sole proprietorship (JDG) with VAT compliance.

What Tripozi is not:

• Not venture-backed (no growth pressure to spam).
• Not a content farm (human editorial review on every vertical; AI limited to low-stakes structured content).
• Not an “SEO affiliate site” auto-generating per-keyword pages we can’t defend.
• Not anonymous: the legal entity is identified on our Privacy and Terms pages, and reader contact is via the contact page.

Full methodology and the four-tier rigour breakdown on the methodology page.

Every personalized itinerary you generate gets a unique URL under /trip/[id] that you can bookmark, share, or email directly. No account required — the trip is stored server-side but indexed by an unguessable ID, so only people with the link can see it.

Additionally, your browser keeps a “recently viewed” list in localStorage, accessible via Your saved. This list never leaves your device.

Account-based saving with PDF export, itinerary history across devices, and multi-trip management is planned for the upcoming Tripozi Pro tier.

We cover 172 destinations spanning 72 countries — every continent except Antarctica. The list grows as we can verify each new destination properly (real GPS coordinates, typical trip length, “popular for” tags, best-months seasonality).

We’d rather describe 172 places carefully than 2,000 places shallowly — that’s exactly the “seems-helpful but isn’t” pattern Google’s Helpful Content Update penalizes.

Accessibility and dietary verticals are limited to 10 European cities (Barcelona, Kraków, Amsterdam, Lisbon, Berlin, Prague, Valencia, Copenhagen, Tallinn, Porto) because that’s where OpenStreetMap has enough contributor-tagged venues to ground the content honestly. We’ll expand as OSM density grows in other cities — not before.

Passport Index and Sherpa are excellent data aggregators — we cross-check against them during our quarterly matrix review.

Tripozi’s difference lives in the source layer: every entry in our 72-combination visa matrix deep-links to the issuing government’s official page (Japan Ministry of Foreign Affairs, Thailand MFA, UAE MOFAIC, Singapore Immigration & Checkpoints Authority, etc.) with a visible last verified date. Aggregators rarely expose their source layer; we do, so you can verify before booking flights.

When we’re less than 100% confident in an answer, we say so explicitly and link you to the primary source and to aggregators for cross-checking — rather than fabricating certainty.

Please do. Three channels:

• Accessibility / dietary venue corrections: each accessibility and dietary city page has a “suggest a venue” form that writes to our moderation queue. Approved submissions are tagged back to OpenStreetMap so the whole open-data ecosystem benefits.
• Visa matrix corrections: email [email protected] with the URL to the official government page showing the correct rule. We review weekly and update quickly — visa rules change and we’d rather learn from a reader than publish stale info.
• Itinerary / tool errors: same email — include the page URL and what’s wrong. Gemini-generated pages can be regenerated on demand.

As little as possible — and we tell you what and why:

• Analytics: self-hosted Plausible on our own infrastructure — no cookies, no cross-site fingerprinting, no personal data. Country-level traffic counts only.
• Saved itineraries: stored as an unguessable ID on our servers but not linked to your identity unless you choose to sign up for the premium tier (not yet live).
• Recently viewed + passport preference: stored in your browser’s localStorage — never leaves your device.
• Ads (when live): Google AdSense via Consent Mode v2. Default DENIED; only activates if you explicitly accept the cookie banner.
• Newsletter: if you subscribe, your email stays on Resend (our email vendor). Unsubscribe links are on every email.

Full breakdown including all data subprocessors (OVH, Cloudflare, Gemini API, OpenStreetMap, Unsplash, Resend, Google AdSense) on the Privacy page.

Google’s Helpful Content Update (HCU) — which killed most large-scale AI-content sites between 2023 and 2026 — targets content that pretends to be helpful but isn’t, regardless of production method. The question isn’t “AI vs human”, it’s “is this actually useful to the person asking the question?”.

Our answer is a tiered trust model:

• High stakes, no AI: visa, accessibility, dietary. Wrong answers have real costs (missed flights, stranded travelers, contaminated meals).
• Medium stakes, AI with heavy guardrails: destination overviews, itinerary sequencing, tool pages (packing, budget, when-to-visit). Gemini output is cached per destination + modifier combo so repeated queries give consistent answers, Redis-backed for 30-60 days, Zod-validated with enum-drift fallbacks.
• Low stakes, pure AI: day-by-day micro-copy (“then walk 5 min to ramen shop”) where verification is on you at the point of use.

We’ve also deindexed ~3,000 pages that would have been weakly-unique (thin combinatorial month pages, duplicate modifier variants) — because depth beats breadth for HCU survival.

Business entity: Polish Jednoosobowa Działalność Gospodarcza (sole proprietorship) with VAT registration. The legal operator is named in our Privacy policy and Terms.

Infrastructure: origin servers in OVH Gravelines, France (EU); Cloudflare handles global edge caching; self-hosted Plausible analytics on same infrastructure; Google Gemini API and Unsplash API as external services.

Applicable law: EU GDPR governs your privacy rights on this site. Polish consumer protection + tax law governs the business. If you’re located in the UK, US, or anywhere else, your local laws still give you consumer protections that we honor — when premium tier launches, Paddle (merchant of record) handles cross-border VAT / sales tax compliance so you pay the right rate for your jurisdiction.

Email [email protected] with the URL and a brief description of what’s wrong. We read everything and fix same-week for verifiable issues.

For visa corrections specifically: please include a link to the official government source (MOFA, MFA, immigration authority, or embassy) showing the correct rule. This lets us update the entry and cite your correction’s source in the next quarterly review.

For accessibility / dietary venue issues: if you can, also file the correction at OpenStreetMap so the whole ecosystem benefits — our data re-syncs from OSM weekly.