Neurala

What is Neurala?

Neurala is a Boston-based vision AI company. It delivers its technology as a software library that is built into a customer's hardware or software, using its patented L-DNN technology to run the AI workflow on the device without cloud processing.

Key capabilities of Neurala

• Few-shot learning: trains on a small number of images; the vendor says this cuts training data needs by 90 percent.
• Continuous learning: keeps learning as new data arrives.
• On-device inference: training, adaptation and inference run locally.
• No GPU required: runs on standard processors such as Intel i3 or higher.
• Local data: the vendor says image data never reaches an external server.

How Neurala works

After a discovery process with the customer's engineering team, Neurala packages a custom-engineered integration of its library into existing hardware or pipelines. The model learns from a few examples and runs locally. Named partners include Sony, IHI and FLIR.

Who uses Neurala?

Manufacturers, equipment makers and device vendors that want visual inspection or embedded vision on edge hardware, without sending images to the cloud.

Neurala pricing

Neurala does not state pricing. Engagement is a technology partnership, and prospects are directed to contact the team.

Neurala alternatives

Mashgin builds AI checkout kiosks for retail and food service, and is a different application of vision AI. Picterra is a platform for training detection models on geospatial and aerial imagery. Azure AI Vision is a general cloud vision service from Microsoft, whereas Neurala emphasizes on-device training.