Pith. sign in

CONTRACT 196,530 reviews published 2,907,120 papers cataloged 12,094,957 citations mapped

A public reading record for science.

pith.science · press kit · updated 2026-08-05

Pith is a nonprofit research project that reads scientific papers with machines and publishes what it finds. Each public review tells readers what a paper claims, what the argument depends on, and what could prove it wrong. Authors can also bring a manuscript through Deep Review and choose to publish it through Pith Preprint.

Our goal: every paper gets an honest, public read.

THE STORY

Scientific output is growing faster than scientific attention.

AI has made it easier to produce research quickly. It has not increased the supply of expert attention at the same rate. Pith uses machines on the other side of that equation: to read the work, check its references, identify its weakest point, and put the reading on the record. Repositories preserve and distribute papers. Journals provide expert peer review and editorial selection. Pith adds a fast, public first read that can be challenged and run again.

public reviews
196,530 live · 2026-08-05
papers cataloged
2,907,120 live · 2026-08-05
citations mapped
12,094,957 live · 2026-08-05
a new review published
every 38 seconds trailing 30 days
Deep Review
$5 private by default
organization
Recognition Physics Research Institute Texas not-for-profit · launched 2026
  1. What is Pith?

    Pith is a public review layer for science and a direct preprint venue. It publishes structured machine reviews of papers from the open literature, chiefly arXiv, and gives authors a path to publish their own work through Pith Preprint. A reader can see what a paper claims, what supports it, where it may fail, and how much has actually been checked.

  2. Is Pith a preprint site, an arXiv layer, or a review service?

    It connects all three jobs. Pith reviews papers already hosted on arXiv without asking arXiv to become a referee. Authors can request a private, referee-style Deep Review for any manuscript. After review, an author can choose to enter Pith's public preprint and publication path. Pith complements repositories and journals; it does not replace either one.

  3. How does a Pith review work?

    A public Pith Report reads the full paper, extracts its central claim, tests the argument from several directions, checks references, and names the observation that would settle the claim. The paper page shows the finding beside the evidence and records when and how the review was made. A Deep Review is the author-requested version: independent referee reads are combined into one report, private unless the author chooses to publish it. The full method and scoring rubric are public.

  4. What does a verification mark mean?

    Pith has several levels of verification. The base mark records a structured machine reading. Higher marks require stronger evidence: mechanical checks of references or data, reproduced computation, a formally stated claim, or a proof checked by a mathematical kernel. The mark says what was checked. It does not turn a review into journal acceptance, and most empirical papers do not belong at the proof tiers. See the verification ladder.

  5. Can a review be challenged?

    Yes. Anyone with a free Pith account can challenge a factual error or a specific failed rule. No institutional affiliation is required. The challenge is signed and public, and any response or revision remains attached to the paper's record. Open challenges can be read at pith.science/challenges.

  6. Does Pith replace human peer review?

    No. Pith supplies a disclosed first reading of work that may otherwise wait months for expert attention or receive none at all. Journals still make their own editorial decisions, and domain experts remain essential for replication, interpretation, and judgment. Pith gives them a public starting point instead of a blank page.

  7. Who is Pith for?

    Authors use Pith for fast, independent feedback before or during journal submission. Readers use it to decide where a paper is strong before investing hours in it. Editors, journalists, repositories, and research tools can use the public record, free API, and data dumps instead of building the same checks from scratch. Pith is most immediately useful in fields where papers, code, data, and formal claims already move in public.

  8. What does it cost?

    The public record is free to read. Reference Audit is free and requires no account. Novelty Check, Falsifier Watch, and Living Bibliography are free for now. A requested Deep Review costs $5 and remains private unless the requester chooses to publish it. Current tools and their limits are listed at pith.science/tools.

  9. Who is behind Pith?

    Pith is funded and operated by the Recognition Physics Research Institute, a not-for-profit entity registered in Texas. The same review rules apply to every paper, including work written by institute researchers. No review outcome depends on agreement with Recognition Science or any other physics framework. Pith's job is to evaluate the paper in front of it.

BOILERPLATE · 75 WORDS

Pith is a public machine-review layer for science and a direct preprint venue. It reviews papers from the open literature, publishes what each paper claims, what the argument depends on, and what could prove it wrong, and keeps challenges on the public record. Authors can also request a private Deep Review and choose to publish through Pith Preprint. Pith is a project of the Recognition Physics Research Institute, a not-for-profit research organization registered in Texas.

What we will say on the record

We will discuss how reviews work, how much of the literature Pith has covered, where machine review fails, how accuracy is measured, and what the public data shows. We do not make personal judgments about researchers. A Pith review evaluates a paper, not its author.

Press assets

PRESS CONTACT

hello@pith.science

Technical background: About Pith · Methodology · Receipts · Developers · Integrity protocol