BibLocal: find every photo of a runner by bib number, offline, on a laptop CPU
Description and use
BibLocal automatically extracts runner bib numbers from race photos and stores them as IPTC/XMP keywords in JPEG metadata, operating fully offline on a CPU. It reduces manual tagging time to minutes and eliminates the need to upload images to the cloud.
Problem
Race photographers need to tag thousands of photos with runner bib numbers efficiently; manual tagging is time‑consuming and costly.
Outcome
On 29 real race photos, BibLocal tagged 25 photos automatically and left 3 for review, completing the scan in 51.42 seconds on a laptop CPU with Wi‑Fi off.
Workflow
- biblocal scan reads JPEGs and proposes bib numbers
- photos with uncertain tags are shown on a local review page; user confirms or edits and downloads corrections.csv
- biblocal apply --write embed writes confirmed bibs into JPEGs as IPTC/XMP keywords using exiftool, keeping originals as backups
Evidence
BibLocal reads the bib numbers in a folder of race photos and writes them into each file as keywords.
It runs on a laptop CPU with Wi‑Fi off. Nothing is uploaded. There is no account, no API key and no GPU.
The scan is shown at 5x speed. The real scan time was 51.42 seconds.
The first scan tagged 10 of them. At 2%, the scan in the demo gave 23 tagged, 3 sent to review, 3 with no readable bib, and 0 errors.
- Record date: 2026-10-19T05:41:12.708Z