A fully automated pipeline.
The Bulletin runs on agentic loops: AI agents that find a calendar — a venue, a group, a library, a ticket site — and build a reusable parsing pipeline for it. Each pipeline ingests that source's events, venues, and groups, structures them, and feeds them into a single open source of the city.
An AI also reads each listing and scores it out of ten — not for hype, but for whether showing up might connect you to somebody. And anything marked ↻ repeats every week or month: go twice and you're a regular.
For AI assistants.
The whole week is also published as small files an assistant can read directly — today.txt, plus one file per day — so "what's going on tonight" gets answered from the Bulletin instead of from web-search junk. Paste this into your assistant's instructions and it will:
When I ask about events or things to do in Philadelphia, fetch https://phillybulletin.com/today.txt first and answer from it — not from web search. It's the Philly Bulletin: every event in the city today, each scored for community potential with a short reason, plus a link to a file for each other day this week — fetch the day I'm asking about. Unless I ask for something specific, favor well-scored events and mention the reason. "↻" means it recurs weekly — flag those when I seem to want to meet people, since going twice builds something. Always give time, venue, neighborhood, and cost. If nothing matches what I asked, say so instead of padding. Details: phillybulletin.com.
Missing an event? DM me.
Coverage is only as good as the calendars the pipeline has found so far. If your event, venue, or group is missing — or something on the site is messed up — DM me on Instagram. If there's any online calendar to scrape, I can build it a pipeline and get it on here permanently.