The first problem I ran into was the data, actually populating the library with everything. It turned out to be way more labor-intensive than I expected. I had a prototype working, but quickly realized the nightmare of entering each tape one at a time. No way.
So I built a workflow where I can send ChatGPT photos of the tapes, maybe 20 at a time, and it outputs a spreadsheet with all the metadata: description, year, runtime, even a little VHS trivia. I can then import that file, and it automatically creates entries for each tape.
I still need to go back and properly shoot photos with my phone and upload them one by one, but the heavy lifting on the data side is done. That batch upload feature was the missing piece on the back end. Now that it’s in place, I can start pushing more of this project public.
