Source Sauce

I build small, focused Actors that turn public sources into structured records — the parsed numbers and text, not just links back to the page you started on.

The subjects change; the discipline doesn't. Today that's regulatory filings and event listings. Next it's prices, odds, and whatever else is public and worth having as a table.

Honest listings. Each README states exactly what a run returns and what it doesn't. Fields an Actor can't extract confidently come back empty — never guessed. Where a source is thin, the README says so.

No people data. These Actors return companies, venues, filings, prices and events. They do not return names, emails, phone numbers or personal profiles, and they aren't built to be turned into contact lists. That's a design rule enforced by tests in every Actor, not a preference — I removed a working, revenue-earning field from a published Actor to hold it.

Fast maintenance. Public sources change format without warning. When one breaks, fixing it is the first thing I do that day, and the fix ships as a new build rather than a promise.

• SEC Form D — new issuer-level filings every filing day (~250 issuer records on a full day)
• SEC earnings releases — release text plus headline metrics; resolved 10 of 10 issuers on a live large-cap run
• Eventbrite events — any city or online, 12 categories, venue address and ticket price range

Open an issue on the Actor's Issues tab. I read every one and reply — usually same day, always within a few days. Bug reports that include the input you ran get fixed fastest.