Sensecap Watcher: Control your smart home by yourself

O Sensecap Watcher is a compact device equipped with a camera, screen and microphone that allows you to control smart home devices such as lamps, electronic devices and appliances. Its technology artificial intelligence allows understanding of events without the need for specific commands.

Created by Seeed Studio, the product is currently seeking crowdfunding on Kickstarter and has already raised around $45,000 from approximately 300 contributors.

Read also
78% of Brazilian startups already use Artificial Intelligence in their processes
This is the first commercial created using Sora, OpenAI’s artificial intelligence

Sensecap Watcher Features

Versatile in its use, Sensecap Watcher can, for example, recognize a delivery man at the door and instruct him where to leave a package or turn off the TV when a mother asks her son to do his homework.

It can also automatically turn off the lights when you fall asleep, scold a dog for making a mess, or turn on a lamp when you pick up a book.

In addition to home use, Watcher has potential for the commercial environment, identifying customers who may need assistance and automatically calling a salesperson.

Setting up the device is simple, with just instructions in natural language to program it according to the user’s needs, such as “let me know if there is no one in the room” or “if the baby cries, tell me mommy is coming”.

AI and sensor compatibility

On the technical side, Sensecap Watcher is compatible with the Home Assistant platform, the IFTTT service and language models from companies such as OpenAI and Meta. It uses both internal AI and cloud-based language models to interpret specific situations.

According to Seeed Studio, the device uses Grove sensors to detect environmental variables such as CO2, humidity and temperature. It can also be connected to Arduino or Raspberry Pi platforms, and the project is open source.

On Kickstarter, Sensecap Watcher is available to anyone who pledges $89. For those concerned about privacy, there is the option to process the language models locally using a Jetson AI computer. Nvidiaalthough this alternative is more expensive: US$ 1,999.

Sources: Kickstarter, Seeed Studio