ReText.AI, with the support of Yandex Cloud, plans to train a neural network to collect and analyze news
ReText.AI service is moving its operations to the Yandex Cloud platform. Through the cloud, synonym checking and text uniqueness assessment are already being performed. However, in the autumn of 2023, ReText.AI will introduce new features: news analysis and punctuation and spelling checking. The beta version of the punctuation checker is already available to users and has demonstrated impressive effectiveness. During internal tests, the grammar checker successfully handled tasks related to punctuation, correctly placing all punctuation marks in sentences with isolated members and in complex sentences.
The development team of the online service ReText.AI made the decision to transition to the Yandex Cloud platform to ensure more stable service operation. Over the past few months, the neural network has evolved from a simple synonymizer into a highly sought-after tool used daily by more than 10,000 users. Working in the cloud allows for flexible resource management based on current load, leading to infrastructure cost optimization and increased project self-sufficiency. Moreover, the transition to the cloud significantly accelerates the process of implementing new functional features into the service.
Yandex Cloud platform already hosts clusters for PostgreSQL, Redis, and virtual machines used in Kubernetes. It's worth noting that the team maintains a 200 GB storage capacity for user data exclusively for the "Synonyms" service, and versions of trained models occupy an additional 100 GB in storage. Additionally, ReText.AI actively utilizes Yandex Object Storage for storing trained models, Docker images, project logs, and database backups.
Currently, the development team of the ReText.AI neural network is focused on creating projects that operate directly in the cloud. One such project is nearly ready for launch—a Telegram bot designed to simplify information field analysis. The bot will analyze available information sources and generate concise summaries of the results. This is just one step in the company's direction toward creating a professional platform that enables specialists to optimize their time, receive high-quality content, and perform in-depth text analysis.