What problem does it solve? Fetching podcast transcripts repeatedly wastes money and time. This Skill polls the All-In Podcast episode list, diffs it against a local SQLite database, and stores each transcript once so episodes can be reused for briefs and cross-episode trend analysis without re-fetching. ## Core Features & Use Cases - Episode polling and diffing: poll_new_episodes.py fetches the episode list with a free HTTP GET and reports only the YouTube video IDs not yet stored. - Idempotent transcript storage: store_transcript.py upserts on video_id, so re-storing an episode never double-charges the paid youtube_transcript scraper. - Database setup and verification: setup_db.py creates the schema idempotently and verify_db.py checks row counts, empty transcripts, and duplicate IDs. - Use Case: A stock-research bot polls for new All-In episodes weekly, fetches only the missing transcripts, stores them in SQLite, and hands them to a distillation skill for investor briefs. ## Quick Start Ask the assistant to set up the All-In transcripts database, poll for new episodes, and store the latest episode transcript.
