AI & Machine Learning
pith.science
Provides tools for machine review and analysis of academic papers, including extracting claims, searching evidence, and generating falsifiers.
ENDPOINT 1
https://pith.science/mcp
MCP server metadata
- Name
- pith
- Version
- 1.1.0
Pith (pith.science) machine-reviews new scientific papers and publishes for each a verdict, novelty and correctness scores, the core claim, the weakest assumption, and a falsifier: the observation that would settle the claim. For a specific paper, route by the requested information: use analyze_paper_claims for what it asserts, its premises, logical dependencies, argument structure, or reasoning; use get_paper_falsifier for a test or result that could disprove it; use get_review for Pith's overall evaluation, scores, and summary. All three accept an arXiv id or URL, DOI, or Pith Number. For a topic or paper title use search_papers; for new reviews in a field use latest_reviews; for who claims something or what evidence exists use search_claims. These tools return Pith's own reviewed record as structured data. For what a paper itself states, under what conditions, with which measurements, and whether its arXiv abstract omits anything its manuscript says, use get_record: the compiled record is the authors' own sentences with typed labels and exists for papers Pith has not reviewed. For work across several papers: paper_packet returns one paper's record compactly, citation_neighbors walks one hop of the citation graph, paper_bridge returns what two papers share, find_passages finds sentences of one role across all stored records, and record_coverage says what was examined; an empty result means nothing in the examined scope matched. Signed human reviews (get_signed_reviews) are structured deltas against the machine review; get_reviewer_record shows a reviewer's calibration.
Known tools 17
get_reviewPith's review of one specific paper, in one call: the verdict, confidence, novelty and correctness scores, one-line summary, weakest assumption, and the paper's argument as short labeled statements (premise, machinery, core claim, falsifier) already extracted, so the paper's premises and logical structure can be stated without reading the PDF.
Inferred read-onlyanalyze_paper_claimsThe structured argument of one specific paper: what it actually asserts, the premises it assumes, the machinery connecting those premises to the claim, and the falsifier, as short labeled statements in argument order (premise, machinery, claim, falsifier).
Inferred read-onlyget_claimsAlias of analyze_paper_claims: the same ledger of what the paper asserts, its premises, machinery, and falsifier, in argument order.
Inferred read-onlysearch_claimsFull-text search over every scientific claim Pith has extracted, across all reviewed papers.
Inferred read-onlyget_paper_falsifierThe falsifier for one specific paper's core claim: the concrete observation, experiment, or computation that would settle or disprove it, plus how many people are watching for that result.
Inferred read-onlyget_signed_reviewsPublished human signed reviews of one paper: reviewer identity, axis deltas against the machine review, claim verdicts, defects, predictions, and who paid.
Inferred read-onlyget_reviewer_recordA reviewer's published signed reviews and, when present, their calibration record (Brier score, open predictions, positivity skew).
Inferred read-onlyget_recordThe compiled record of one paper: the authors' own sentences, arranged with typed labels.
Inferred read-onlyresolve_paperTurn a pasted arXiv id, arXiv URL, DOI, or Pith Number into the arXiv id the other tools key on.
Inferred read-onlypaper_packetOne paper's compiled record, small enough to hold beside others: the main claim and contribution in the authors' words, each abstract sentence with its role and probability, paper-level labels, and, for full-text records, conditions, measurements, results and limitations, with totals so a capped list is not mistaken for the whole paper.
Inferred read-onlycitation_neighborsOne hop from a paper: every reference entry with its resolved work identity, the papers that cite it, and, where a record exists, the citing sentences with their labeled intent.
Inferred read-onlypaper_bridgeWhether either paper cites the other, the resolved works both cite (on or off arXiv), and the papers that cite both.
Inferred read-onlyfind_passagesAuthor sentences Jev labeled with one role, across every stored record, each with its probability and paper.
Inferred read-onlyrecord_coverageSources read and levels run for one paper's record, or the number of stored records by tier.
Inferred read-onlyCONNECT WITH APPROVAL
Client installation
Review this server and its permissions before adding it. Secret placeholders must be set locally.
Codex
~/.codex/config.toml
[mcp_servers.pith]
url = "https://pith.science/mcp"
enabled = trueClaude Code
.mcp.json
{
"mcpServers": {
"pith": {
"type": "http",
"url": "https://pith.science/mcp"
}
}
}Claude Desktop
Settings → Connectors → Add custom connector
Name: pith
Remote MCP URL: https://pith.science/mcpAdd this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.
Cursor
.cursor/mcp.json
{
"mcpServers": {
"pith": {
"url": "https://pith.science/mcp"
}
}
}Visual Studio Code
.vscode/mcp.json
Add to Visual Studio Code{
"servers": {
"pith": {
"type": "http",
"url": "https://pith.science/mcp"
}
}
}Generic MCP
Client-specific MCP configuration
{
"name": "pith",
"transport": "streamable-http",
"url": "https://pith.science/mcp"
}MCP Inspector
Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.
TRUST AND VERIFICATION EVIDENCE
Loading Trust v2 evidence…
Checking the associated registrable domain. The BuiltWith key remains server-side.
Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.