Mary Lou

I build fast, reliable scrapers for ads, marketing and e-commerce data — turning public ad libraries, shops and search platforms into clean, analysis-ready datasets.

🎯 What I work on

Ad libraries & ad intelligence:

• Meta (Facebook & Instagram) Ad Library — copy, CTAs, image and HD video URLs
• Google Ads Transparency Center — Search, YouTube, Display, Shopping, Maps
• TikTok Commercial Content Library — 32 EU/EEA markets, downloadable creatives
• LinkedIn Ad Library — who paid, impressions by country, targeting
• TikTok Creative Center — top ads with CTR, likes and landing pages

E-commerce & lead intelligence:

• TikTok Shop — prices, units sold, shops, variants and stock
• LinkedIn Jobs — no login, no cookie, no account at risk
• Google Search — organic SERP results, built for rank tracking
• Website Contacts — emails, phones and socials from any domain list

💸 Pricing: from $0.14 per 1,000 results. No subscriptions, no start fees.

🛠 How I build

Python · direct API/GraphQL over browser automation · TLS fingerprinting · resumable checkpointed jobs

• Measure every transport constant, never copy one. chrome146 works on Meta and is refused outright by Google.
• No browser until it's proven necessary. Headless Google Search took 75s/query against 1.3s over plain HTTP, so I deleted it.
• Runs that survive. Global dedup, checkpoints, wind-down before the timeout, a hard budget cap, and a second pass for refused targets. A blocked run still ends successfully and says why.
• Input read the way it was meant: any link format or field name, never a failed run.
• A live view in every run, with results so far.
• Schemas that admit what they don't know — if a field is empty upstream or derived rather than read, the docs say so.