Exhibit H · The automation stack
A scraping engine, a real-time relay, an archiver and a ranked twice-daily briefing. All working, all private, because they hold live sessions and keys.
These carry authenticated browser sessions, API keys and personal data. Publishing them would mean either shipping credentials or gutting the parts that make them work. So they are described here rather than linked, and labelled private rather than left vaguely implied.
A universal scraper. Renders with an undetected browser and extracts structured data with a model, for the pages that return a block page or an empty shell to anything simpler. It reads sites that defeat plain fetching, including LinkedIn.
A Discord to WhatsApp relay. Watches a trading community and reformats its calls into structured tables, in real time, in the place I actually read messages. Planned setups stay in their own section rather than being mixed in with live calls, because conflating the two is how you act on something that was never a signal.
An Instagram story archiver. Stories expire in 24 hours, so anything worth keeping has to be captured automatically or it is gone. Captures with review notes attached.
A briefing pipeline. Ranks everything I am tracking and pushes a top five to WhatsApp twice a day on a launchd schedule. The ranking is the hard part: capture is easy, deciding what deserves attention is not.
A dashboard I have to remember to open is a dashboard I will not open. The briefing goes to WhatsApp because that is where I already look, and the dashboard exists for when I want depth rather than a nudge.
An unranked feed is a second inbox, which is a cost rather than a tool. Five items, ordered, twice a day, is a decision aid.