I build Actors that handle the unglamorous half of web data work: the part that happens after the scraping is done.

Scrapers are good at collecting. They are rarely good at producing something you can hand to a client, load into a CRM, or run analysis on. Field names arrive in three different casings, empty values show up as "N/A" instead of null, and the same entity appears four times under slightly different spellings.

That gap is what I work on.

Small, composable Actors that do one job well and chain cleanly onto whatever scraper you are already running. No lock-in, no proprietary formats — dataset in, dataset out.

Current focus areas:

• Normalization — consistent field names, whitespace, numbers, emails, phone formats
• Deduplication — exact matching, and fuzzy matching for near-duplicate entities
• Validation — dropping incomplete records before they reach your database

Clear pricing. Pay-per-event, charged on output, not on input. If an Actor removes 40% of your rows as duplicates, you pay for the 60% you keep.

Honest documentation. Every Actor README states what the tool does not handle and where its limits are. You should know that before you run it, not after.