From El Jardín to Ember: The Aittendee Experience
summary
Live Ember app captures audio in ~12‑second chunks, transcribes with Whisper → transcript chunks, summarizes into chapters, extracts definitions and pull quotes via GPT‑4 prompts, generates illustrations with an art‑director model and image API, and voices historical attendee questions with Google text‑to‑speech. Pace meter reports 152 wpm over 14.4 minutes, sentiment and subjectivity score each chapter, and per‑API accounting totals $4.72 total. Architecture and code are open on GitHub as a working Ember example of Aittendee.
AI-written summary · about
chapters (6)
- 0:00Introduction and background
- 5:46Why intelligent applications matter
- 10:20The Garden of Earthly Delights overview
- 21:27Live AI demo introduction
- 27:32How the AI app works
- 35:00Demo recap and closing
