How Vecto makes Deep Learning Search Available for the Regular Developer

David Ooi
29 Sept 2022

The latest in AI technology, transformer models, are able to represent human language and images in a way that preserves the meaning inside them. By utilizing teams of data scientists, innovative companies are able to use these models to enable searching by semantic meaning (for both images and documents) using natural language. Semantic search helps customers find the things they are looking or quicker, uplifting sales significantly.

We created Vecto to bring that same innovation to every company. 

Vecto enables you to perform the same deep learning-based search using a simple REST API. No AI or machine learning experience required.

If you have a lot of unlabeled images, unstructured documents, and have no way of quickly finding the images or documents at the time of need. Vecto has you covered.

Vecto enables you to index all your unlabeled and unstructured images and documents, and you will be able to search for those images and documents using only natural language such as English. Vecto understands the semantics of images, it understands English and is able to relate between images and English, this allows Vecto to retrieve relevant images and documents using only English as a search query.

This is also known as Vector Search, aka Deep Learning Search. It enables text to image query, image to text query, text to text query and image to image query.
Vecto has another amazing feature called Analogy Search, we will talk about in an upcoming blogpost.

Vecto’s simple API is perfect for any regular developer without any AI background. You can use Vecto’s API by simply performing HTTP POST requests, which can be done in any programming languages of your choice, such as JavaScript, Python, C++, or any softwares that can send POST request such as Postman, or directly via the Linux curl command line tool.

Vecto’s simple API is also perfect for any kind of search related tasks. For example, if you are running an online store, you could index your products’ images and description in Vecto, this provides better search results for your customer at your online store.

If you have an existing SQL database with many rows of improperly labeled images or documents, you could index every single one of those rows in Vecto, this allows you to indirectly query your database via Vecto using natural language such as English.

Vecto’s simple API is extremely flexible for many use cases, and it is easy to learn and use. Take a look here for your first tutorial on Vecto! Have fun!

Vecto has another amazing feature called Analogy Search, we will talk about in Part 2.