1000usdinchina.com Dev Retrospective (1) - China Travel Budget Estimator: 100 Cities, Built Solo

How much does a trip to China actually cost? That question has almost no good answer online. You can find flight prices and a hotel here and there, but nobody tells a first-time visitor what the whole thing costs end to end. So I built 1000usdinchina.com — a free China travel budget estimator that covers 100 cities, and I shipped it solo using an AI coding agent.

This is post 1 of an 8-part series on how the site was designed, built, tested, and shipped. This one is the why: the problem, the market, and the product. The next seven go deep on the engineering.

Table of contents

• The problem: nobody knows what a China trip costs
• Why now: the visa-free window
• What the product does
• The business case
• The stack, in one sentence
• What the rest of this series covers
• FAQ

The problem: nobody knows what a China trip costs

A traveler planning two weeks in China faces a wall of unknowns. Is ¥1,000/night normal for a hotel in Chengdu or a rip-off? How much is a bowl of noodles in Xi'an versus a sit-down dinner in Shanghai? What does the high-speed train from Beijing to Qingdao cost, and how long does it take? Each answer lives on a different site, half of them behind a login, most of them in Chinese.

The result: people either over-save and never go, or under-budget and get a nasty surprise. A budget estimator that answers "can I do China on $1,000?" in one screen didn't exist. So that became the product.

Why now: the visa-free window

Timing matters for a travel product. China has rapidly expanded visa-free and transit-visa-free entry for dozens of nationalities. Suddenly a huge pool of would-be visitors can go with far less friction — and the first thing they search is some version of "how much does a China trip cost." Building the answer to that query, in their language, at the moment the barrier dropped, is the whole bet.

What the product does

You give it a few inputs and it gives you a costed trip:

Budget:   $1,000
Entry:    Beijing
Exit:     Shanghai
Month:    October
Travelers: 1 adult

It returns a route through real cities, a per-night and per-meal breakdown, transport between cities, and a total you can compare against your budget — localized into the traveler's own language and currency.

flowchart LR
    A[Budget + entry/exit + month + travelers] --> B{Estimator}
    B --> C[Suggested route across real cities]
    B --> D[Lodging / food / transit / attractions cost]
    C --> E[Total vs. your budget]
    D --> E
    E --> F[Shareable trip card in your language + currency]

Two things make the numbers trustworthy:

1. Real, curated data — hotel, food, transit and attraction figures for 100 cities, cleaned from large source datasets (covered in post 3 on the ETL pipeline).
2. Language ↔ currency coupling — pick English and you see USD; Japanese, JPY; Korean, KRW. Zero-decimal currencies are rounded correctly. No half-yen prices.

The business case

A free estimator isn't charity — it's the top of a funnel:

• Demand is huge and intent is high. "China trip cost" is a query typed by people about to spend money on flights and hotels. That's the most valuable kind of traffic.
• The data is a moat. 100 cities of cleaned, aggregated, compliant travel data isn't something a competitor clones in a weekend.
• A monetization path exists. Affiliate links for booking, premium itineraries, and partnerships are a natural V2.0 — without ever charging the traveler to estimate.

And the punchline for indie hackers: this was built by one person. Not because I'm fast, but because I ran an AI coding agent (Claude Code) under a strict set of engineering disciplines. That's the subject of the flagship post 5.

The stack, in one sentence

Next.js 15 on Cloudflare Pages, with D1 (SQLite at the edge) for dynamic data, Workers for compute, an AI trip-planner on Cloudflare Workers AI, and a fully static-rendered core for speed — all glued together and shipped by a single developer plus an agent. Each piece gets its own post below.

What the rest of this series covers

Key takeaways

• A high-intent, underserved search query ("how much does a China trip cost") + a timing tailwind (visa-free expansion) is a strong product wedge.
• Curated, compliant, multi-city data is a defensible moat for a travel tool.
• One developer + a disciplined AI coding agent can ship a 100-city app end to end.

FAQ

How much does a trip to China cost? It depends on cities, season, and travel style, but many routes are doable around $1,000 for a week or two. 1000usdinchina.com estimates a real total from your budget, entry/exit cities, month, and number of travelers.

Is the estimator free? Yes. Estimating a trip is free and requires no account.

How many cities does it cover? 100 Chinese cities, with curated lodging, food, transit, and attraction data.

What languages and currencies are supported? Multiple languages, each paired with the traveler's local currency (e.g. English/USD, Japanese/JPY, Korean/KRW), with correct rounding for zero-decimal currencies.

Next in the series → Running a travel app fully on the edge: Next.js 15 + Cloudflare Pages, D1, Workers