How to build AI apps with Vectara

Vectara has recently brought the development of productive artificial intelligence applications to a new level by introducing the Vectara portal. A pioneer of RAG, the Palo Alto, California-based company has designed an open-text environment to enable anyone to build AI programs to interact with their data.

Vectara portal

Vectara has distinguished itself with ease of access and use. Unlike many commercial products that help users get instant answers from their documents and data, this portal allows users to build their own data-driven search, summarization, or chat applications in just a few simple steps, without even having to write One line of code!

This tool has a lot of potential for non-programmers and can be used in various cases, including policy making and invoice search. However, since the tool is still in beta and only a limited number of customers are testing it, its performance has yet to be fully evaluated.

Ofer Mendelevitch, Vectara’s head of developer relations, said that by supporting and offering the company’s proprietary RAG platform as a service, they expect to see widespread adoption by non-programmers. “We look forward to seeing what users will build with the Vectara portal and hope the accuracy and relevance enriched by their documentation will showcase the full power of Vectara’s enterprise RAG systems,” he added.

How the Vectara portal works

Available as a Vectara-hosted application as well as an open source product under the Apache 2.0 license, Vectara Portal allows users to easily build and share their own custom applications for interacting with data.

To get started, users need to create a portal account using their main Vectara account and set it up with information such as Vectara ID, API Key, and OAuth client ID. After setting the profile, users can create their new portal by clicking the create portal button and entering the basic details such as program name, description and program function type and add it to the portal management page.

From the portal management page, users can open the created portal and change its settings. They can upload the necessary documents to contextualize and customize the application. Once uploaded, these documents are indexed by the platform to provide accurate answers without illusions.

This platform, which includes a powerful retrieval engine, an advanced model, and a multilingual retriever, dramatically increases the quality of responses overall. Using this no-code product, creators can quickly create their own productive AI products with just a few clicks.

The platform also allows users to choose from different language models, including company-specific models and OpenAI LLMs, so that their application answers questions in the most optimal way possible.

Enterprise customers

With this no-code product, offered as both a hosted and open source product, Vectara seeks to expand the reach of enterprise users and provide the ability to build powerful, productive AI applications. The company’s goal is to increase signups and build buzz for its core RAG-as-a-Service product, ultimately turning it into a useful tool for organizations.