OCR | z.AI

OCR | z.AI is a free web-based optical character recognition tool powered by GLM-OCR, a lightweight 0.9B-parameter multimodal model developed by Z.ai (formerly Zhipu AI). It extracts text from images and scanned documents with state-of-the-art accuracy, achieving a score of 94.62 on OmniDocBench V1.5, and supports complex layouts including tables, formulas, handwriting, and multilingual text across 100+ languages. The tool outputs structured formats such as Markdown, JSON, and LaTeX, and also supports structured information extraction from invoices, receipts, and forms via JSON schema.