Vectara

Formerly Zir-AI

Vectara is a website and app search platform powered by large language models (LLMs), according to the vendor, Vectara, Inc. It utilizes natural language processing and AI to deliver search results and enhance user experience. The product is designed to serve enterprises and organizations of all sizes. It caters to various professions and industries such as IT professionals, marketers, developers, customer support teams, research and analysis professionals, e-commerce companies, financial institutions, legal firms, healthcare organizations, and educational institutions.

Key Features

Blazing Fast Semantic Search: According to the vendor, Vectara provides AI native search for websites and apps with a P50 of 60ms. It includes snippet extraction and cross-attentional AI re-ranking for efficient semantic search.

Breakthrough Search Relevance: Vectara focuses on the idea rather than the words, using natural language processing and zero-shot models, according to the vendor. It aims to deliver higher recall and precision for user queries and applies Neural2 reranking for improved search results.

Vectara Neural Rank™ Model for Semantic Search: According to the vendor, Vectara utilizes large language models and zero-shot model techniques to achieve a broad understanding regardless of the question or language used. It eliminates the need for keyword-based techniques and offers language support.

Document Search: Vectara processes various document formats, including product manuals, support articles, question and answer databases, website pages, emails, blogs, financial reports, and legal documents, according to the vendor. It generates vector-based relationships and provides highlighted pointers to relevant ideas within documents.

Text Analysis: According to the vendor, Vectara applies advanced AI models to analyze text from multiple sources such as emails, social media comments, web pages, customer support tickets, messages, survey responses, tweets, and documents. It aims to assist with recommendations, identifying related content, generating Q&A responses, and query routing.

AI that Obeys the Rules: Vectara's AI search understands meaning and context, according to the vendor. It allows customization of search results based on custom business rules and actions. The vendor claims it supports dynamic filtering, cross-corpora searching, and related content visibility.

Suggested Responses: According to the vendor, Vectara empowers sites, applications, and customer support teams to respond promptly and accurately to user questions. It delivers what the vendor claims are the most accurate responses from across the breadth of organizational data.

APIs: Vectara offers a range of options for data ingestion, including a no-code admin interface and a complete library of APIs, according to the vendor. These APIs are designed for application developers and data engineers, abstracting away the complexity of operating an ML search pipeline.

Custom Dimensions: According to the vendor, Vectara supports custom dimensions that directly influence search results. Developers can attach key data attributes, such as star ratings, product categories, document freshness, or custom probability scores, to order results before serving them to users.