Example use cases

Each use case below has a full walkthrough on olaverselab.com, with the pipeline step by step, benchmark figures from the model cards, and code.

Decisions and extraction

Automated decisionsRoute, flag and score tickets, chats and records with calibrated probabilities, in English and eight African languages.

Document extractionFill a JSON template from a photo, a scan or plain text, with a calibrated confidence on every field.

Payment messagesPull the amount, sender, reference and date out of mobile-money and bank SMS, then decide whether it is genuine.

Language and search

Language routingDetect the language of tickets, chats, queries and documents across 608 languages, before you act on them.

Diacritic restorationRestore tone marks on text typed or scraped without them, from Yorùbá and Igbo to Vietnamese and Polish.

Arabic diacritizationRestore Arabic tashkīl, with or without case endings, for text-to-speech, teaching and readable copy.

Search that matchesRestore lost marks, then match on an accent-folded key, so typing “oko” finds “ọkọ”.

Questions and training data

Question generationNatural, search-style questions from any passage in 25 languages, for quizzes, FAQs and assessment banks.

Retrieval training dataGenerate verified (query, passage) pairs from your own corpus to fine-tune retrievers and rerankers.

Low-resource corpus miningSort web crawls into 608 languages on CPUs and pull out the African and low-resource text.