Tool not found.
Back to Home
Fast, open-source Rust library that classifies PDFs as text-based or scanned and extracts clean, structured Markdown without OCR, running locally in under 200ms.
pdf-inspector is a Rust library built by Firecrawl to handle text-based PDFs locally, skipping expensive OCR services for the majority of PDFs that don't need them. It performs smart classification, detecting whether a document is TextBased, Scanned, ImageBased, or Mixed within milliseconds by sampling content streams, and returns a confidence score along with per-page OCR routing recommendations. The library provides position-aware text extraction with font info and automatic multi-column reading order, then converts documents into clean Markdown, handling headings, lists, code blocks, tables, bold/italic formatting, and URL linking. It uses dual-mode table detection, combining rectangle-based detection from PDF drawing operations with heuristic detection from text alignment, and supports CID font decoding for a wide range of encodings. The document is parsed once and shared between detection and extraction, avoiding redundant work, and the entire parser can run as WebAssembly directly in browsers and Web Workers with no server round trip. It's a lightweight, pure-Rust library with no ML models or external service dependencies, and ships with bindings for Python, Node.js, and the browser, alongside CLI tools for converting PDFs to Markdown from the command line.
