Pith: peer-to-peer review and an open, living scientific record
No gates.
Send us a paper and it gets read on its merits. The rules are public, and every author gets the same ones.
Reference check: free, with an account. Full review: $5, private until you choose to publish. Half of the last 100 reviews finished within 29 minutes of being sent, 9 in 10 within 66 minutes (measured 15 Oct to 15 Nov 2026).
Read the film
Thank you for submitting your manuscript. Your research is important to us. Please continue to wait. Current wait: fourteen months. The gates exist because the record had to be right. And there weren't enough reviewers to handle them all. A single company makes a billion dollars in profit off volunteer expert reviews. We regret to inform you. No system is beyond repair. Please... remain in line. There's a better way. Pith. Peer-to-peer reviews and an open, living scientific record.
Start free. Publish when the paper is ready.
-
Check your references
Free · with an account
Every reference is looked up in Crossref, DataCite, OpenAlex and arXiv and checked against retraction notices. Each problem is quoted, and Pith lists the report nowhere.
-
Get a full review
$5 · half within 29 min, 9 in 10 within 66 min (last 100 reviews)
Pith reads the whole paper and writes what it claims, what those claims rest on, where it could break, and what to fix. The review stays private unless you publish it.
-
Publish it
Whenever you choose
The paper goes up on Pith with its review beside it, free, whatever the review concludes. For $10 Pith also puts it on the record: a DOI for a paper that is not on arXiv, and the exact PDF anchored in Bitcoin.
-
Ask an expert
From $200 · 7 to 14 days
A signed review by a listed reviewer in the paper's field, for a paper on arXiv or on Pith. It is published whatever it concludes.
A Pith Review is written by a model and can be wrong, so each one is versioned, signed and open to appeal. A signed review is written by a person, and the reviewer is paid for it. That is why Pith offers both.
One set of rules, and everyone can read them.
- No sponsor
No endorser and no institutional email. Sign in with a verified email and you can submit.
- No score bar
Any paper can be published with its review, whatever the score. The score and any blocking finding sit beside the paper for every reader to weigh. The rubric reads the review's findings, never the author's name or institution.
- No private veto
An editor can act only on an integrity hold, and every such decision goes into a public log.
- In public
Publication counts and score distributions are public on the audit page. Anyone signed in can challenge a claim, with evidence, on a public ledger.
498,316
papers reviewed
live · read 15 Nov 2026, 01:18 UTC
38,279,161
citations mapped
live · read 15 Nov 2026, 01:09 UTC
Referees get paid.
Scientists spent more than 100 million hours refereeing in 2020 alone, unpaid.1 On Pith the same expertise is paid work.
- Offers
We send papers that match your fields. You accept or decline. Nothing is assigned over your head.
- Signed
You write the review on the paper and sign it with your name or a pen name. It is published, and your record grows with every review.
- Paid
Stripe's card fee comes off the buyer's payment first, Pith keeps 15% of what is left, and you receive the rest.
- Same either way
A negative review pays the same as a positive one, and a review never decides whether the paper is published.
Reading is what a gate stands in for.
Science needs someone to check a paper before the world builds on it. Journals did that with expert reviewers, and there were never enough of them.
So papers wait, and the reviewing is mostly unpaid. Preprint servers were supposed to be the open door. arXiv, which was that door for most of its 35 years, now asks a first-time author for an institutional email and prior authorship in the field, or a personal endorsement from an established author.5
A system that can read every paper, check every reference and publish the evidence beside the paper doesn't have to ask who you are first. That is the gate Pith takes away.
Built for readers that check.
More people now meet science through an AI assistant, and an assistant has to decide which sources to believe. A journal's logo tells it very little. Evidence tells it a lot.
Every Pith review is published as structured data. A read-only server speaks the Model Context Protocol, so an assistant can ask for a paper's review, its claims, its signed reviews, a reviewer's track record or the papers around it in the citation graph, and get the evidence back along with the judgment. An AI that can check doesn't have to take anyone's word for it, ours included.
- llms.txt
A map of Pith for assistants, with the MCP server address.
- Bulk downloads
Every current review, in one file a month.
- API
The Pith API: papers, reviews and verdicts as JSON.
- Accuracy
How the reviews were measured, and what they missed.