A clearer view of the work.

A local workspace for finding, reviewing and selecting images and films.

My contribution
Workflow design, curation and AI-assisted implementation
Built with
Python / FFmpeg / File hashing / Local web UI

The problem

Thousands of images and films were spread across working folders. Reviewing one file at a time made it difficult to choose a coherent portfolio.

The build

Scan metadata, compare file hashes, generate thumbnails and sampled video frames, and keep review decisions in a local catalogue. Prepare separate web copies of the selected assets.

The result

The inventory contains 5,591 files and identifies 1,481 exact duplicate copies. A separate editor manages portfolio copy, cover choices, order and visibility. Original media remains untouched.

Inspect the work

The local workflow compares file hashes for exact duplicates, creates previews and stores review decisions. Selected media is prepared separately for the portfolio.

The homepage demonstration uses only the 16 selected portfolio assets. The full file archive, private paths and review decisions stay on the local machine.