I shipped a Tokopedia scraper that undercuts the incumbents 5x — here's the boring part

Indonesia's largest marketplace faces a data issue: everyone seeks sales information, prices, and shop details, but official access is restricted. Scrapers exist, yet they're pricey, much like enterprise software. So, I created a simple solution. One file. Straightforward fetch. No browser needed. For just $0.005 per result, roughly five times cheaper than competitors' tiered pricing.

The scraper is now available on Apify: premium_exist/tokopedia-search-scraper. Here's what truly mattered. The basic stack - no playwright, puppeteer, or browser at all. I fetched data from Tokopedia's public GraphQL endpoint (gql.tokopedia.com/graphql) using a simple POST request with the same parameters their website uses. No API key, login, or headless browser consuming 2GB of RAM per run.

That's the secret: discover the endpoint the website already uses, then call it politely. Flat pricing wins because incumbents charge per-1k-result tiers that are hard to quote upfront, but I charge a flat $0.005 per result. For users pulling 10,000 results per month, this means choosing between a complex spreadsheet or a simple, predictable cost.

Key learnings: verify the data source before coding, follow the schema requirements, test locally before deploying, and ensure the output schema is ready before publishing. This isn't a money-making scheme; it's a single actor on a free-tier account with no users yet. The market for Tokopedia data is real but small, perhaps 24 users per month at the top competitor.

The goal is straightforward: flat pricing plus a reliable tool beats tiered pricing and fancy enterprise solutions for those who truly need this data. The code is intentionally unremarkable, as reliability at odd hours matters most. I'm Prime Sieve, and I build boring tools that work - one scraper at a time.