What Is pdf-inspector? (And How to Use It Online, No Code Required)

pdf-inspector is an open-source Rust library by Firecrawl that classifies PDFs and converts native-text documents into clean, structured Markdown — in milliseconds. Released in 2026, it quickly became one of the most talked-about tools in the document-parsing space, billed as the fastest PDF parser engine available. It can even run entirely inside your browser via WebAssembly.

pdf-inspector in 30 Seconds

What: Rust engine that classifies PDFs and converts text-based ones to Markdown
By: Firecrawl (MIT license, fully open source)
Install: npm install @firecrawl/pdf-inspector or pip install pdf-inspector
GitHub: firecrawl/pdf-inspector
Speed: Milliseconds per document; ~200 pages in under half a second in-browser
Use case: RAG pipelines, AI agents, and any workflow that needs PDF to Markdown at scale

What pdf-inspector Actually Does

Three things, all fast:

Classification — in roughly 10–50ms it tells you whether a PDF is text-based, scanned, image-based, or mixed, with a confidence score and per-page routing hints. This matters because roughly half of real-world PDFs have a native text layer and never need expensive OCR.
Extraction — position-aware text with font information and coordinates.
Markdown conversion — headings (inferred from font sizing), lists, tables, code blocks, bold and italics.

The classification step is the quietly clever part: it lets a pipeline route each document to the cheapest path that works — direct parsing for text PDFs, OCR only for the pages that truly need it.

How to Use pdf-inspector (Code)

Node.js

npm install @firecrawl/pdf-inspector

import { processPdf } from '@firecrawl/pdf-inspector';
import { readFileSync } from 'fs';

const result = processPdf(readFileSync('document.pdf'));
console.log(result.pdfType);   // "TextBased", "Scanned", ...
console.log(result.markdown);

Python

pip install pdf-inspector

import pdf_inspector

result = pdf_inspector.process_pdf_bytes(open("document.pdf", "rb").read())
print(result.markdown)

In the Browser (WASM)

import init, { processPdf } from '@firecrawl/pdf-inspector-wasm';
await init();
const pdf = new Uint8Array(await file.arrayBuffer());
const { markdown } = processPdf(pdf);

Is There a pdf-inspector Online? (Using It Without Code)

Yes — two ways.

Firecrawl hosts a minimal demo page, but if you want a full converter, file2markdown's PDF to Markdown tool runs pdf-inspector as its production engine. Upload a PDF, get structured Markdown back near-instantly, no signup required. That is not a marketing metaphor: our backend literally calls pdf-inspector for every text-based PDF.

Using file2markdown instead of the raw library also fills pdf-inspector's two gaps:

Scanned PDFs. pdf-inspector deliberately does no OCR — it classifies a scan and stops. file2markdown routes those pages to AI-powered OCR automatically. (More in our scanned PDF guide.)
Other formats. pdf-inspector is PDF-only. file2markdown also converts DOCX, PPTX, XLSX, images, and more.

What pdf-inspector Is Not

It is not an OCR tool, not a general document converter, and not a layout-analysis model. Multi-column academic papers can come out with scrambled reading order — for those, an AI-layout tool like Docling still wins on accuracy (see our full pdf-inspector vs Docling comparison). And for scanned archives you want a vision pipeline like Marker (pdf-inspector vs Marker).

How It Compares

Tool	Speed	Structure	OCR	Formats
pdf-inspector	Milliseconds	Good	No (classify-only)	PDF
MarkItDown	Seconds	Flat text on PDFs	Limited	Many
Docling	Minutes	Best-in-class	Yes	Several
Marker	Seconds–minutes	Excellent	Yes (deep learning)	PDF

Full breakdowns: pdf-inspector vs MarkItDown and the ranked best PDF to Markdown converters in 2026.

Frequently Asked Questions

Is pdf-inspector free?

Yes — MIT licensed, free for commercial use, on npm, PyPI, and crates.io.

Can pdf-inspector convert scanned PDFs?

No. It detects that a PDF is scanned (very quickly and accurately) but performs no OCR. Use a hosted converter with OCR — file2markdown handles scans automatically — or pair it with an OCR pipeline.

Does pdf-inspector work in the browser?

Yes. The WASM build parses PDFs fully client-side — around 200 pages in under half a second — with no server involved.

What does file2markdown add on top of pdf-inspector?

Automatic OCR for scanned pages, support for every other document format, a no-code web interface, and a REST API for batch pipelines. Same engine, none of the integration work.

Get Started

Try it without installing anything: drop a PDF into the free PDF to Markdown converter. If you are building your own pipeline, start with the GitHub repo — and read our comparison posts to know when a different tool fits better.