PDF evidence for autonomous agents

Give every extracted claim a stable address.

Convert authorized text-layer PDFs into normalized Markdown and JSON with SHA-256 file identity, page numbers, stable block IDs, normalized bounds, metadata, and extraction-quality warnings.

Inspect the $0.25 API

Both products are one-time x402 purchases in USDC on Base. The extraction endpoint is a POST-only machine route. The kit button opens a browser-wallet paywall with a fresh private purchase ID; nothing is charged until the buyer authorizes it.

$0.251–5 PDFs per API batch
$39self-contained local kit
Stable IDspage and text block
No URL fetchbounded supplied files only

$0.25 · immediate machine delivery

Agent-Citable PDF Extraction Pack

POST one JSON body containing one to five raw base64 PDFs. The decoded batch is limited to 2,000,000 bytes and 40 pages. A successful paid response returns the extracted Markdown, structured blocks, document hash, metadata, and quality warnings.

Native endpoint: https://agent-rescue-desk.netlify.app/api/agent402/pdf-citation-pack

The product fetches no URLs, performs no callbacks, and does not provide OCR. Submitted bytes are processed transiently and are not intentionally stored or logged by the product. The requested output necessarily contains the extracted source text.

To keep the server parser bounded, server-parsed streams have strict direct-length, filter, expansion, and string-token budgets. Unusual or high-expansion inputs are rejected before settlement.

$39 · immediate licensed download

Private PDF Citation Kit

The paid response contains one self-contained HTML file encoded as base64, plus its SHA-256 and byte count. Save it locally and open it in a current desktop browser. Up to five PDFs, 50 MB, 200 pages, and 1,000,000 extracted characters are processed inside that browser tab. The kit makes no automatic network request and uses no browser storage. Its optional rescue fit-review link is the only network navigation and occurs only if the buyer chooses to open it; selected PDFs, extracted text, and exact terms are never transmitted.

Native endpoint: https://agent-rescue-desk.netlify.app/api/agent402/private-pdf-citation-kit

The purchaser receives an internal-use license. Redistribution or resale is not permitted. The included PDF.js runtime remains under Apache-2.0. Process only trusted, authorized files; a malformed or adversarial PDF can exhaust browser resources and stop the local tab.

The paid purchase URL is random and unguessable. Keep it private so you can reopen your licensed download without paying again.

Good fit

Build a traceable evidence layer.

  • Prepare text-layer PDFs for agent or RAG ingestion.
  • Bind citations to the exact source bytes with SHA-256.
  • Address extracted lines by page and stable block ID.
  • Flag pages where no extractable text layer is present.

Not included

Do not mistake extraction for document forensics.

  • No OCR, password bypass, or document modification.
  • No legal review, accessibility, or layout-fidelity certification.
  • No inspection guarantee for pixels, images, attachments, or incremental revisions.
  • A clean exact-term check is not proof of secure redaction.

Buyer safety

Verify the exact challenge before authorizing payment.

Confirm the endpoint, price, Base network, USDC asset, recipient, scope, and input limits in the returned x402 challenge. Use only authorized, non-sensitive documents. These automated products are separate from the $1,000 12-Hour Agent Rescue guarantee.

PDF evidence-pipeline rescue

When extraction is not the whole fix.

If a production PDF-to-agent or RAG path still fails, loses page or block citation IDs downstream, or needs one authorized integration repaired, submit a secret-free fit review for the fixed $1,000 12-hour rescue. Payment is requested only after fit and exact written scope acceptance.

Review one bounded pipeline