Images, scanned PDFs, and foreign-language documents — full-text recognition preserves layout, and non-Chinese content is automatically translated into Chinese in a streaming manner; chart photos can also be directly restored into exportable data tables.
Recognition results retain the original row-column relationships, and columns in table-type documents are aligned without misplacement.
Automatic detection of English/Japanese/Korean/Russian/Arabic and other languages, with segmented streaming translation during recognition and bilingual comparison.
Original image box selection on the left, structured text on the right, bidirectional linked highlighting, doubling verification efficiency.
Photos of line charts/bar charts are directly extracted into data tables and redrawn, and analysts can ask follow-up questions about trends in a conversational manner.