A post about something visual is weaker without the visual, and until this week getting one into Pluma meant me manually screenshotting and uploading after the fact.
Now any Claude Code session, on any of my projects, can render a real image straight from the actual design artifact or a genuine Simulator screenshot, upload it, and stage it, without writing a word of the post itself. Pluma’s own Scout, Strategist, and Editor still decide whether an image actually fits a given week’s theme and write the caption. I built it this way on purpose: letting a session freelance both the picture and the copy would skip every rule the Editor enforces, character limits, app name placement, currency formatting, and that’s exactly the inconsistency this whole exercise was meant to fix.
Each staged image now carries its real app icon baked into the frame, not just a text label, because I’m building enough different things at once that a follower needs to recognize which app a picture is about without reading the caption first.
