📄PDF to Text
Fetch a PDF (by URL, SSRF-guarded) or accept it as base64 and extract all readable text with page count. Deterministic parsing — no language model, so nothing is invented. Text-based PDFs only; scanned images need OCR (not included).
🔌 Server-side tool
This tool runs on the server (it fetches third-party URLs directly and cannot run in your browser). Use the paid JSON API below — the worked examples show real server-side results and refresh automatically.
Worked examples
Live examples refresh automatically — call the API for guaranteed-fresh results.
Machine API (x402)
$0.003 / callThis tool is also a JSON API for AI agents. Requests without payment receive 402 Payment Required plus instructions; agents pay USDC on Base via the x402 protocol — no accounts, no API keys.
GET /api/v1/dev/pdf-to-text?url=https%3A%2F%2Fwww.w3.org%2FWAI%2FER%2Ftests%2Fxhtml%2Ftestfiles%2Fresources%2Fpdf%2Fdummy.pdf HTTP/1.1
Host: agenttools-hub.vercel.app
→ 402 (payment required, instructions in headers)
→ 200 (after X-PAYMENT header; JSON body below)
{
"tool": "dev/pdf-to-text",
"input": {"url":"https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf"},
"result": { "value": null, "answer": "" }
}Agent docs: /llms.txt · OpenAPI spec · integration guide
About this tool
Language models read text, not PDFs. This tool extracts the full text layer from any PDF — by URL or base64 — with page counts and hard size caps, server-side and SSRF-guarded. Deterministic extraction: the output is exactly what the document contains, nothing invented.
Frequently asked questions
Does it work on scanned documents?
No — scanned PDFs are images and need OCR. The tool detects this and returns a clear error instead of empty text.
What are the limits?
10MB per file, 150 pages, and a configurable text cap (default 50k characters, extendable to 200k). Fetches are SSRF-guarded.