Tuned to your game list

Recognition is sharpest when it knows what to expect. When you choose a region, TrailCamHub loads the game species that occur there — and then hands the list to you.

You stay in control

• Remove species you'll never see at your spot.
• Add others manually if your ground is unusual.
• Mark the ones you consider most likely.

How your list shapes results

The species you keep form an allow-list for that camera's region. When the AI analyses a photo, its result is checked against this list:

• A recognised species that's on your list is reported as-is.
• A result that isn't on your list is flagged for a second look rather than shown as fact — these are the cases most likely to be a mistake.
• Species you marked as most likely act as a gentle tie-breaker when the model is unsure between similar animals.

Crucially, your list speaks the model's own language: each species is mapped to the exact label the recognition model uses internally — and the European and global models name things differently — so the matching is reliable, not approximate.

What it is not

This doesn't retrain the model on your photos. It's a relevance-and-review layer on top of recognition: it narrows the field of plausible answers and surfaces the doubtful ones, which keeps results trustworthy without any machine-learning work on your side.

The result is fewer odd misclassifications, and clear flags on the handful of detections genuinely worth your judgement.