TrapTagger
Open-source AI web app for processing and classifying camera-trap images
About
TrapTagger is a free web application from the nonprofit WildEye Conservation that runs camera-trap images and videos through three AI models: detection and blank filtering, species classification, and coat-pattern matching for individual identification. It was built with Oxford's WildCRU on a dataset of over one million annotated images, with a classifier covering 55 Southern African species, and it also flags humans and vehicles. The code is open-source on GitHub, a free hosted account is available, and uncertain detections are left for manual annotation rather than auto-labeled.
What it can do
- AI detection and blank-image filtering
- Species classification (55 Southern African species)
- Coat-pattern individual identification via HotSpotter integration
- Manual annotation workflows for uncertain images
- CSV/spreadsheet export of detection data
Setup
- Sign up for a free account at traptagger.co.uk or self-host from GitHub
- Upload camera-trap images or videos to your project
- Let the AI pipeline detect animals and filter empty frames
- Review automatic species labels and manually annotate uncertain cases
- Download a CSV or spreadsheet of detections by species, location, and date
Prerequisites
Camera-trap image or video sets; browser access (or a server for self-hosting)
Use cases
Works with
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: Verified via official wildeyeconservation.org pages and the open-source GitHub repo today.; Free hosted account plus self-hosting; 55-species Southern Africa classifier trained with Oxford WildCRU on 1M+ annotated images.; Classifier coverage is Southern Africa-focused; other regions need custom training. GitHub star count not recorded.. License: Open-source; nonprofit-hosted free tier; uncertain detections are left for manual annotation..
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 6.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 9.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · Verified via official wildeyeconservation.org pages and the open-source GitHub repo today.
- · Free hosted account plus self-hosting; 55-species Southern Africa classifier trained with Oxford WildCRU on 1M+ annotated images.
- · Classifier coverage is Southern Africa-focused; other regions need custom training. GitHub star count not recorded.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.