Vibeleaderboard
← Back to Tools
Index / tool

pdf-inspector

github.com/firecrawl/pdf-inspector
Visit github.com
Category
Developer Tools
Rank
No. 1993Tools index

Previous survey · No. 2037 ·

Pricing
Open Source
Type
TOOL
Use case
Data Processing
Interfaces
CLI
Builder
@nickscamara_
GitHub
19.6k stars
Latest release
v1.25.2
Date

About

A Rust library for PDF inspection, classification, and text extraction. It detects whether a PDF is text-based or scanned so callers can make smart routing decisions, extracts text with position awareness, and converts documents to clean Markdown without OCR, with selective OCR available to native Rust and CLI consumers. Bindings are provided for Python and Node.

What it does

Firecrawl's project page presents pdf-inspector as a gate in front of OCR: it samples a PDF's content streams for text operators, sorts the file into one of four types, and lists which pages need OCR. Firecrawl says it built this to keep text PDFs local and out of paid OCR services.

Stated on the product site

License
MIT
Rust install
Published as a crate, added with cargo
Python install
Installed from pip
Node and browser packages
Separate npm packages for Node.js and browser WebAssembly
Command-line tools
Two binaries, pdf2md for conversion and detect-pdf for classification only
OCR runtime
PDFium and ONNX Runtime must be installed separately when pages are routed to OCR
Default classification sample
Eight evenly spread pages unless another scan strategy is chosen

Not stated on the site

  • The benchmark table is the project's own run on 200 PDFs with OCR disabled; the page shows no independent result on scanned or mixed files.
  • The figure that about 54% of PDFs need no OCR comes without a stated corpus or method on the page.
  • The page does not show how classification confidence behaves on PDFs with broken font encodings beyond flagging them for OCR fallback.
  • Platform support for the OCR build is deferred to a separate setup guide that the page does not reproduce.

Written from the product site at github.com.

Tags

pdfrusttext-extractionmarkdownocrlibrary

Tech Stack

PythonRust

Comments (0)

No comments yet

Editorially curated, with community endorsements as a secondary signal. Corrections welcome.