Emil Stetco

FLAIR: detects sarcasm, satire, and irony through NLP. Romanian language.

Determines the polarity of any text content: negative, neutral, positive (cross-language).

Detects the language of a given piece of text (170 languages) and offers more language information.

An engine that automatically identifies and labels objects in images.

Speech to Text (Cross-Language).

vLLM-based local LLM engine for generation, embeddings, vision, and JSON output

An engine for multilingual structured text and layout extraction from images and PDFs (vllm).

Finds key moments in a given video and turns them into a simple engaging summary for any viewer.

Translates a given text into a target language.

Named Entity extraction with Relations from multi-language (cross language) content.

Zetta Assistant Platform uses Zetta Local LLM to act as your personal assistant.

Determines the polarity of any text content: negative, neutral, positive. The Romanian version.

Extracts all public metadata for an article specified by URL.

Extracts toponyms from text and assigns geolocation data automatically (Romanian language).

Extracts article plain text from html page (URL or uploaded html file)

Romanian authorship attribution engine identifying likely author from input text.

ZETTA dataset upload service for ClearML-based NER and classification training workflows.

Converts Romanian numerals and validates standard codes (CNP, IBAN, CAEN, CUI).

Transcribes (OCR) text from an image (.jpg, .png, .tif, .pdf) - Multi-lang text recognition model.

Transcribes text out of given file (.jpg, .png, .tif, .pdf) using a Romanian text recognition model.

Automatic document organization into clusters of similar documents.