Replicate 推出 Datalab Marker 與 OCR 模型：快速將文件與圖片轉換為 Markdown 與精確文字定位
Original: Extract text from documents and images with Datalab Marker and OCR
The Replicate platform has newly listed two powerful document and image parsing models developed by Datalab: "Datalab Marker" and "Datalab…
Replicate 平台上架了來自 Datalab 的兩款全新文件解析模型：Marker 與 OCR。Marker 專為將整份複雜文件（如 PDF）轉換為乾淨的 Markdown 格式而設計，非常適合 RAG 應用；OCR 模型則能精確提取圖片或文件中的文字，並提供行級（line-level）的多邊形定位座標，為開發者提供高效的文件預處理方案。
The Replicate platform has newly listed two powerful document and image parsing models developed by Datalab: "Datalab Marker" and "Datalab OCR." They are designed to address the pain points developers face when extracting and structuring text from PDFs, scanned files, and images. In today's era of booming generative AI and RAG (Retrieval-Augmented Generation) applications, converting unstructured PDF documents into clean, machine-readable text remains one of the greatest challenges developers encounter.
Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.
See Pro plans →Want the original English / full article?
Read on Replicate Blog →Summaries are AI-generated; the original article is authoritative.