Modules
Six specialized engines, one pipeline
From a raw scan to a grounded answer, each module handles one step of the pipeline. Here is what each engine does and when it comes into play.
- MIKA
MIKA: multi-modal document analysis
MIKA reads text, tables, images and scanned pages together and turns mixed-format documents into a single structured knowledge layer.
Read more → - MOSAIC
MOSAIC: multi-observer analysis with intelligent consensus
MOSAIC runs several specialized models over the same content and reconciles their output into one consensus result with fewer errors.
Read more → - VERA
VERA: verified extraction with region anchoring
VERA anchors every extracted value to its exact location in the source document, so any answer can be traced and verified in one click.
Read more → - OCR
OCR: from scans and images to searchable text
RootNia's OCR module turns scanned pages, PDFs and images into searchable, machine-readable text, ready to classify and query.
Read more → - RAG
RAG: answers grounded in your own documentation
What RAG (retrieval-augmented generation) is and how RootNia uses it so every LLM answer is grounded in your own documents.
Read more → - NL-to-SQL
NL-to-SQL: query databases in plain language
NL-to-SQL turns plain-language questions into SQL queries so anyone on the team can query structured databases without writing code.
Read more →