Monitor bees with AI

See through the hive to change lives

How does it work?

Count bees

Measure precise colony size and predict potential swarming

Model beehives

It's difficult to observe and record all colony information using pen and paper while in the field. Model hives to remember where to take action and why

Classify beecomb

Take a photo of honeycombs to get detection of cells and ratio of resources. Balancing nectar and pollen between colonies can prevent starvation

beehive frame with detected cell types


Finding a queen in a bustling metropolis to determine if the colony has a future can be quite challenging. We can surprise you and find two


Take snapshot of the beehive state to track colony development over time

AI beekeeper

Given all of the hive context, get a one-button advice from a large language model on the next steps

beehive frame with detected cell types


Gather context of plants and weather around your apiary. Use satellite maps to know better pollinateable areas.

Plantnet integration Tallinn Tallinn

Varroa mites (beta)

Detect and count varroa mites infesting your bees without the need of killing bees with alcohol washing test

Entrance video monitoring (beta)

Stream entrance video to get detected bees coming in and out depending on time

Why Gratheon

Honeybees are important for natural pollination and food security. To hobby beekeepers, this super-organism is also a cute pet.

Traditional beekeeping involves hard physical labor and frequent inspections to keep track of the colony state. Monitoring is needed so that colonies survive varroa mite infestation, starvation, hornet attacks, robbing state, late swarming, viral and bacterial infections and pesticide exposure.

Inspections are limited in time and need full attention to spot smallest details. Manual inspections do not scale well with the amount of colonies and growth of the vertical hives.

Without proper technological solutions, most beekeepers to this day rely on using alcohol washing tests that kill hundreds of bees just to get level of mite infestation.

We help with observability so that you can be a better beekeeper

Shape the future of beekeeping

Our long-term vision is a robotic beehive. Interested in details how it would work?

Gratheon app is fully open source. We provide flexible GraphQL API with API tokens to import or export data.

We're looking for partners and investors to speed up and scale our development. We are looking for engineers to join our team. We are looking for beekeepers to test our product.

Join us. Got ideas to share or questions to ask?

Join in Discord Sign up


Gratheon is built on the shoulders of giants. We pick and integrate the best open-source models, that are often backed by serious scientific publications.

Our frame resource detection is based on Automatic detection and classification of honey bee comb cells using deep learning by Thiago S. Alves, M. Alice Pinto, Paulo Ventura, Cátia J. Neves, David G. Biron, Arnaldo C. Junior, Pedro L. De Paula Filho, Pedro J. Rodrigues,

Our bee detection uses YoLoV5 model trained on dataset by Matt Nudi

Our gate counter model is based on work done by Fabian Hickert for his Master thesis