A per-download paper marketplace

Every read pays the researcher who wrote it.

Scholar402 charges a few cents in stablecoin every time a human — or an AI agent — opens the full text of a paper. Split automatically: 80% to the authors, 10% to the reviewers, 10% keeps the lights on. No subscriptions, no publisher middleman.

$ GET /api/v1/papers/9f2e.../full
Host: api.scholar402.com
 
HTTP/1.1 402 Payment Required
PAYMENT-REQUIRED: eyJzY2hlbWUiOiJl...
── agent signs a $0.25 USDC payment ──
GET /api/v1/papers/9f2e.../full
PAYMENT-SIGNATURE: eyJwYXlsb2FkIjp7...
 
HTTP/1.1 200 OK
PAYMENT-RESPONSE: eyJzdWNjZXNzIjp0...
Content-Type: application/pdf
author $0.20 reviewer $0.025 platform $0.025

The problem

Three parties, none of them paid.

01 — The author

Writes the paper, sees none of the access revenue

Publishers charge $30–$50 per download. The person who did the research sees exactly $0 of it.

02 — The reviewer

Does unpaid labor that makes the paper trustworthy

Peer review is the whole reason a paper is worth reading — and the only academic labor market where the going rate is zero.

03 — The agent

Has no clean way to pay for what it reads

AI agents are becoming real consumers of paper content — but there's no account to create, no card to charge, no consent flow that works headless.

The mechanism

Paywall enforced by a status code, not a login.

01

Upload, price, split

A researcher uploads a PDF, sets a price between $0.10–$0.50, and assigns a revenue split across every co-author.

02

Abstract is free, full text is a 402

Anyone — human or agent — can read the abstract for nothing. Requesting the full PDF without payment returns an HTTP 402, with the price and payment address in the response.

03

The agent pays and retries

An x402-compatible client signs a stablecoin payment, attaches it to the retried request, and gets the PDF back — no human in the loop, no account, no card.

04

The split settles automatically

Every transaction is logged and swept into a scheduled payout run — co-authors by their configured share, reviewers evenly, no manual accounting.

80% authors
10%
10%
Authors split by configured share, not just evenly — a paper with three co-authors can weight credit however they agreed.

Reviewer and platform shares are computed the same way and recorded per-transaction — the ledger, not a spreadsheet, is the source of truth for who's owed what.

Infrastructure

Built to run lean before it runs big.

Scholar402 runs entirely on Cloudflare's edge platform and the open x402 payment protocol — no servers to keep warm, no idle compute cost between downloads. The whole stack scales to zero when nobody's reading, and scales out globally the moment they are.

WorkersEdge compute, the API itself
D1Ledger & paper metadata
R2PDF storage, zero egress fees
QueuesAsync transaction recording
WorkflowsDurable payout batching
x402Open HTTP payment protocol

For partners & funders

Early, deliberately small, and looking for non-dilutive support first.

Where things stand

Scholar402 is pre-launch. The payment mechanism, revenue-split ledger, and payout engine are built and being verified end-to-end on testnet before a single euro is spent on growth.

This page exists to find early researchers willing to list a paper, agent builders who want a paper-access tool their agents can actually pay for, and funders who'd rather back something already built than a slide deck.

What we're looking for

  • Non-dilutive grants first — infrastructure and protocol-research funding fits this project's actual shape better than equity at this stage.
  • Researchers willing to be among the first papers listed.
  • Feedback from anyone building AI agents that read and cite papers.

Get word when it opens.

No spam, no drip campaign — one email when Scholar402 is ready for its first real papers.

Researcher, agent builder, or funder — say which in the email and you'll hear back directly.