We Made PDF Parsing 100x Faster For Claude Code

pdf-inspector is an open source PDF parser from Firecrawl that lets coding agents process PDFs without waiting on OCR, classifying any PDF in around 20ms and extracting clean markdown locally. Written in Rust with no AI models needed, pdf-inspector reads the content stream, samples pages to classify a document as text-based, scanned, or mixed, then handles glyph extraction, layout ordering, and table detection before converting everything to clean markdown. It processes 200 PDFs in 2.8 seconds and ranks among the fastest runs on the open data loader benchmark. In this video I walk through how PDF parsing and PDF to markdown conversion actually work under the hood, run the WASM build in the browser, and show how to wire it into agents like Claude Code through the CLI or one of the language SDKs. Useful if you're building RAG pipelines, document extraction workflows, or local PDF processing for AI agents. Use the coupon FIRECRAWLYT for 500 extra credits pdf-inspector - https://firecrawl.github.io/pdf-inspector/ Firecrawl anydoc - https://github.com/firecrawl/anydoc 👇 Learn more about Firecrawl 👇 🔥 Website: https://www.firecrawl.dev?via=youtube 📚 Docs: https://docs.firecrawl.dev/features/parse 📱 More Links: 💻 GitHub: https://github.com/firecrawl/firecrawl 🐦 X (Twitter): https://x.com/firecrawl_dev 💼 Linkedin: https://linkedin.com/company/firecrawl