Terrific.ai gives you confidence that your production process is up and running properly

Terrific.AI consists of 2 parts:

A hardware kit for installation on the production line

Mobile app for managers

Terrific.AI differs from traditional vision systems in its unique "learning by observing" approach. Just install the camera, and the system automatically starts collecting and processing images of the manufactured products.

As soon as enough data is collected, Terrific.AI trains a neural network that is able to reconstruct the expected shapes of objects and compare them against the observed images coming from the camera.

If the difference between the reconstructed image and the observation is significant, the system reports the anomaly to an operator. Human feedback is used for improving the neural network, which works towards building a system that is capable of fully-automated quality control.

Trusted by

Our processes are 100% aligned with the strategy of our partners

Mobile App Features

LIVE STREAM
AND COUNTS

Monitor production output in real time, see if it is running or not, what is produced right now, review product photos

PRODUCTION
TRAIL

See all events through the day

ALERTS
CENTER

Setup alerts for line start, stop, or any anomalies

QUALITY
INSPECTOR

Production output history search and comparison. Forward images to your team or client

PRODUCTIVITY BOOSTER

Check the amount of product for different periods, compare shifts and identify bottlenecks

Case Study – Beverages

Detection of wrinkles and dents on aluminum beverage cans

Situation

Get the line stopped immediately when product defects occur. Packing beverages in aluminum cans is a high-speed process. The packaging line may start to generate damages, and the product loss can rapidly accumulate. Identifying such situations and reacting to them as quickly as possible reduces potential losses.

Solution

01

Terrific.AI supports high-speed production (up to 30 items per second)

02

Works in tough production environments - in high humidity, moisture, washdowns

03

Alarms are triggered by serial defects instead of occasional single anomalies - preventing big loss without unnecessary delays to normal production

04

It is non-invasive – doesn't require any serious changes to the production equipment or process

05

Smart AI filters distinguish defects from normal noise (water drops, dust, reflections) – less false alerts

06

Learns new products as they are introduced on the line (self-supervised learning)

07

Works with various defects - doesn't require manual data labeling to start detecting anomalies

Value

Beverage packaging process may generate more than 1% of defected cans. Stopping serial defects will save you tons of product and decrease client complaints.