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.
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.
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.
402PAY
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.