AI‑CLAW

Services

Four types of work. Each can stand alone or combine into a pipeline that keeps running. I write the code myself and hand back cleaned, structured data — not a pile of raw HTML.

Web Data Extraction

Public web data pulled at scale. We agree on the field structure first, then I run the full job — you don’t get columns you can’t use.

  • Good for: store locators, e-commerce catalogs, directories, public social content
  • Delivered as: CSV / JSON / Excel, cleaned and deduplicated, with a field sheet
  • Data I’ve pulled: Google Maps and Yahoo Japan store records, e-commerce storefront catalogs

Browser Automation

Repetitive browser work, scripted away. Chrome DevTools Protocol and Playwright based — handles login, pagination, dynamic content and common anti-bot flows short of CAPTCHA.

  • Good for: back-office systems with no API, bulk report exports, form filling, JS-rendered pages
  • Delivered as: a reusable script with setup notes, or I host and run it on a schedule
  • Stack: Node.js / Python + Playwright + CDP

Lead & Contact Lists

Targeted B2B contact discovery by keyword, industry and country — with email and phone where publicly available. Every row carries its source URL so you can verify it.

  • Good for: overseas distributors, repair shops, store lists, industry buyer contacts
  • Delivered as: structured table, deduplicated, with source link, keyword and collection date
  • Coverage: targeted by country and city; non-English keywords supported (Spanish, Japanese)

Custom Data Pipelines

Scheduled collection, storage and delivery. Your dataset refreshes itself instead of going stale — built for ongoing price, stock, directory or competitor tracking.

  • Good for: price monitoring, competitor catalog tracking, directory refreshes
  • Delivered as: scheduled job with incremental updates and failure alerts, landing where you want it
  • Cadence: daily, weekly or monthly — your call

Where I draw the line

Publicly available data only. I don’t bypass paywalls, log into accounts I’m not authorized for, or access private systems. If a source’s terms prohibit collection, I tell you before we start rather than after.

Login-gated content, sources with legal exposure, and personal private data — I pass on those.

How to start

  1. Tell me what you need. Target sites, fields, volume, format.
  2. Fixed price within 24 hours. Timeline included. Walk away if it doesn’t suit you.
  3. Sample run. You approve a small batch before I scale up.
Request a quote About me first