wtf is legislative-tracker?

Legislative Tracker is a tool that watches state and federal legislation on your behalf so nobody on your team has to read raw bills. Every day it checks for newly introduced or updated bills, uses an AI model to turn each one into a short plain language summary anyone can read in under a minute, sorts it by topic, and sends it to whichever teams need to know about it. The tool pulls bill data from public legislative data services covering all fifty states plus the US Congress. Each bill gets an AI written brief covering what the bill does, who it affects, its current status, and why it matters, written in everyday language instead of legal wording. Bills are tagged by topic and routed, and a single bill can go to more than one team if it is relevant to several. Anything the system cannot confidently match to a topic goes into a review queue instead of being dropped. Summaries can be delivered several ways at once: a daily email digest per team, messages posted into Microsoft Teams or Slack channels, records created directly inside Salesforce with team based routing so each team gets its own filtered list and reports, or plain text file digests for a simple setup with no extra infrastructure. If one delivery method fails, the others still work. The tool also remembers bills it has already seen so it can flag when a bill's status changes, such as moving from introduced to passed, without repeating the same bill twice. The system is built so its pieces can be swapped: the AI model connection uses a common, widely supported format so switching providers is a small configuration change, and the legislative data source is similarly isolated so a different provider could be added later. Since bill data is public record, there is no personal or sensitive information involved, which the author notes makes this easier to approve in organizations with strict data rules, such as government or education. It runs inside Docker so nothing needs to be installed directly on your computer, and can be scheduled to run automatically each day. The author describes this as a portfolio project and states the code is not licensed for open reuse, so anyone wanting commercial use would need to contact the author directly.