EthoStudio is where studies are designed, video is annotated, and models are trained on your species - in your context, by the people who know them best.
The species you prioritise need continuous, structured behavioural data - for welfare, husbandry, compliance. Your keepers produce it by hand, in the hours they can carve out from everything else. Off-the-shelf computer vision can’t fill the gap. It hasn’t seen your animals, your enclosures, or the way your team tracks the behaviours that matter.
EthoStudio works the other way around. Instead of asking your zoo to fit a generic AI, your keepers teach an AI to fit your zoo.

EthoStudio walks your team through six steps, each one built around the way behavioural research is actually done.
Confirm the study and start scoring.
Pick the species you're studying. Search by common or scientific name. EthoStudio will suggest a sampling method that suits the species you've chosen.
Choose how you'll observe. EthoStudio offers four classical methods - focal-animal, scan, all-occurrence and ad-libitum - and recommends the one that suits your species and study length.
Name the individual animals you'll be watching. Each gets a colour, an ID and any notes your team needs. As your keepers tag individuals across video, EthoStudio learns to recognise them by sight - work that pays off when the trained model is deployed on Antz Edge.
Give the study a title, a date range, and a schedule. Note your research goals and anything else the team should know.
Build the list of behaviours you want to track. Each one gets a name, a 3-letter code for fast scoring, and a colour. Behaviours can have sub-behaviours and parameters. Save the ethogram as a template and reuse it across studies.
Confirm the study and start scoring.
Pick the species you're studying. Search by common or scientific name. EthoStudio will suggest a sampling method that suits the species you've chosen.
Choose how you'll observe. EthoStudio offers four classical methods - focal-animal, scan, all-occurrence and ad-libitum - and recommends the one that suits your species and study length.
EthoStudio’s labelling interface is where most of your team’s time will be spent. It’s designed for the way keepers actually work - fast, hands-on, and low-friction.
Narrate what you see, and EthoStudio writes the ethogram for you. Your keepers speak normally - “Deano is foraging sweet potato in his left cheek as Dellah enters the enclosure” - and the system records each animal, each behaviour, and each timed event automatically.
As your model improves, EthoStudio uses it to pre-label new video for your team: so they correct what’s wrong instead of labelling from scratch. The more behaviour your keepers teach the system, the less labelling they have to do.
Use video annotation to label uploaded footage frame by frame, or live observation to score behaviours in the field as they happen - pen, paper and stopwatch replaced. Both feed the same study record.
Video, audio and images all sit in one place. Audio comes with waveform and spectrogram views for bioacoustic work, so a behaviour and its sound can be scored together.
Every behaviour your keepers label adds to what the AI knows. As lessons add up, EthoStudio trains a custom behaviour-recognition model - a piece of software that can recognise the same behaviours on new video, automatically.
The model only knows what your keepers have taught it. It sees the species you prioritise, doing the behaviours your team cares about, in the enclosures you actually have.
The more your team labels, the more it can recognise on its own - until new footage can be analysed without anyone sitting in front of it.

EthoStudio is where your team teaches the AI. Antz Edge is where the AI starts watching - on your premises, in your enclosures, around the clock.
Book a walkthrough. We'll show you how your keepers design a study, label video, and train a behaviour-recognition model on your own species - then watch it run on Antz Edge.
Book a Demo© Copyright 2026 Antz Systems. All Rights Reserved.