Data Mill

Messy pages in, clean JSON out.

I build scrapers for sources that are awkward on purpose: legacy encodings, undocumented internal APIs, search flows that only work in one language, sites that block anything running in a datacenter. Those are the ones worth doing, because they are the ones nobody else has bothered with.

• 🧱 Reliability over cleverness — if a source blocks datacenter IPs, the Actor ships with the right proxy already switched on. You should not have to discover that yourself
• ✅ Verified, not assumed — sold prices are confirmed sold, odds come from official endpoints, transcripts come from the API that still works
• 🌍 English output, any source — Japanese, multilingual, whatever the origin. Clean English keys downstream
• 🤖 Agent-ready — usable over API and MCP, so your agents can pull data on demand
• 🛠️ Maintained — sites change. Open an issue and it usually gets fixed within a day
• 📬 Need a source that doesn't exist yet? Open an issue on any Actor. Custom scrapers built on request

---

Actors

• 🎬 YouTube Transcript Scraper — bulk transcripts from videos, playlists, channels and searches. Segments, plain text, SRT, VTT and RAG chunks
• 💴 Mercari Sold Price Scraper — verified sold comps across Mercari, Yahoo Auctions and PayPay Flea Market
• 🏨 Japan Hotel Scraper — Jalan rates, ratings and plans, including ryokan the global OTAs never list
• 🏇 Japan Horse Racing Scraper — JRA race cards, live odds and results with full payouts