Inspiration

I wanted an easier way to share a set of photos with the story behind them. A photo can show a gathering or someone giving a speech, but it can’t tell you why people came together or what was said. Those are often the details I want to remember and share.

What it does

Photo to Story is a Skill for UGREEN’s AI assistant, Uliya. It uses selected photos and details from the owner to draft a story they can review and share. If an important detail is missing, it asks one focused question rather than guessing. The owner can keep the draft, add context, or correct it. Nothing gets posted automatically.

How I built it

I wrote the workflow in SKILL.md and designed it to use Uliya’s photo expert for read-only photo discovery and inspection. The Skill keeps track of which photos it could inspect and which details the owner confirmed. I rehearsed the workflow off-device and ran UGREEN’s static checker on the submission package. Testing on the NAS is still ahead.

Challenges and what I learned

The hardest part was knowing when to ask a question. An early version asked for too much up front. I changed it to use what the owner has already provided, draft as soon as there’s enough context, and ask only when an answer would change the story. I also planned for partial access and mixed albums: if a photo can’t be inspected or two events can’t be linked, the Skill should say so rather than fill in the blanks.

Built With

Share this project:

Updates

Submission history