Paremiology — the science of studying proverbs and sayings — has long remained scattered among printed books, scans, and closed databases. Senior Engineer Dmytro Yemelyanov decided to fix this and created an open corpus of Ukrainian proverbs verba. The site contains 48,787 expressions covering 185 years of paremiography — from 1841 to 2009.
Paremiology — the science of studying proverbs and sayings — has long remained scattered among printed books, scans, and closed databases. Senior Engineer Dmytro Yemelyanov decided to fix this and created an open corpus of Ukrainian proverbs verba. The site contains 48,787 expressions covering 185 years of paremiography — from 1841 to 2009.
verba is a large-scale digital array that combines 48,787 records of Ukrainian paremiography, compiled into a single unified corpus of five fundamental historical sources. Chronologically, the materials cover the period from 1841 to the present day.
The foundation of the platform was the digitized collections of prominent figures and institutions:
Optical text recognition using Tesseract technology was used to transfer these historical publications into digital format, and the material from the Nomis collection was additionally processed through the publication's critical apparatus.
Each individual entry in the corpus is transformed into a structured card suitable for search, in-depth research, and reuse. It contains the original text with full preservation of the historical spelling of the source, its standardized modern spelling, an accurate reference to the original source, and a group of variants that highlights the connections between similar dialectal or orthographic counterparts.
The adaptation of texts to modern spelling and their thematic classification were performed using artificial intelligence: the accuracy of AI-generated modern spelling is approximately 95%, and the relevance of identified topics is about 85%.
You can search for proverbs by specific keyword, selected topic, or primary source.
However, the main feature of the platform is semantic search - the "by content" function. It is indispensable in cases where the user does not remember the exact words of a proverb, but only its general idea or essence. Artificial intelligence recognizes the context of the query and selects folk sayings that are as close in meaning as possible.
Each expression has a thematic markup that divides folk art into 27 fixed categories, and one entry can be assigned from one to three topics at a time.
For those who have visited the site for inspiration or want to get to know Ukrainian paremiography better, there is a “Browse” mode — it allows you to explore the database in a free format. All the pearls of folk wisdom you find that you like can be added to “Saved” for quick access in the future.
In addition to the web version, the project has a convenient mobile extension — the official Telegram bot ( @VerbaCorpus ). It duplicates the main functionality of the site, allowing you to search directly in the messenger window.
The IT guy talked about the technical stack on his LinkedIn: "The corpus is accompanied by a full-fledged REST API based on Cloudflare Workers, semantic and lexical search, a generator of interactive GIF/PNG cards for distribution, as well as open formats (JSON, XML, CSV, Croissant) and a CC BY 4.0 license."
You can learn more about the project's code and data on GitHub .
Автошкола — це місце, де формуються впевнені навички безпечного керування, і avtokursi-dosaaf.kh.ua є чудовим прикладом якісної підготовки майбутніх водіїв. Навчання поєднує ґрунтовну теоретичну базу, практичні заняття з досвідченими інструкторами та сучасний автопарк. Індивідуальний підхід до кожного учня, комфортні умови й професійний супровід допомагають успішно опанувати водіння та впевнено підготуватися до складання іспитів.