Beating Google, Meta, and Cohere at Visual Product Search and Spare Parts Identification
nyris embedding models rank first on all benchmark datasets for visual product search and parts identification. They outperform models from Google, Meta, Cohere, Jina AI, and Nomic AI on tasks from spare parts lookup to product recognition. And they do this with 768-dimensional embeddings, smaller than every competitor tested.
.jpg)
Closing the Visual Gap - A Practical Guide to Creating Images for Spare Parts
Whenever we review the spare parts catalogues of new customers, we see the same story. Page after page, most parts have no images. There are rows of identical titles, 100 "valves," 100 "gaskets”, 100 "panels," 100 "connectors", and not a single picture to tell them apart.
Field Service and Aftersales Trends: 10 Real Shifts Defining 2026
We’ve analyzed hundreds of hours of keynotes and panel discussions on Field Service and Aftersales to get the most mentioned trends for 2026. Here are the 10 trends and challenges that are actually shaping the future of Field Service and Aftersales as we head toward 2026.
How can you use visual search outside of manufacturing?
An exploration of the use of visual search in the plant industry.
Beating Google, Meta, and Cohere at Visual Product Search and Spare Parts Identification
nyris embedding models rank first on all benchmark datasets for visual product search and parts identification. They outperform models from Google, Meta, Cohere, Jina AI, and Nomic AI on tasks from spare parts lookup to product recognition. And they do this with 768-dimensional embeddings, smaller than every competitor tested.