Can a prompt replace Scribe?

Build a browser extension that turns a recorded workflow into a step-by-step guide automatically. Use a Chrome extension (Manifest V3) that, once recording is toggled on, listens for click events on the page and, for each click, captures a screenshot of the visible tab and a basic description of the clicked element — its visible text or aria-label, wrapped in a generic template like 'Click [element text]'. Assemble the captured steps into an ordered list with screenshots, editable afterward — the user can rewrite any auto-generated instruction, reorder, or delete steps. Draw a simple highlight box on each screenshot around the clicked element, with a setting for cropped or full-page capture. Export the finished guide as a self-contained HTML file or PDF with embedded images. Store guides locally in the browser's storage, with a simple library view. Do not build desktop-app capture beyond the browser, a hosted team library with sharing links, or auto-generated wording more sophisticated than the basic template described — polishing that language is genuinely hard natural-language work, and this build is upfront that it produces a rough draft, not a finished guide. No API key or account needed.