API for agents
A REST API, a Model Context Protocol server and Markdown versions of every page. Discovery is free; premium research costs $0.20 per request via x402.
Quickstart
1. Find something (free):
curl "https://space-kitty.com/api/v1/search?q=surface+code+threshold&type=all&limit=5"2. Inspect a dossier before buying (free):
curl "https://space-kitty.com/api/v1/dossiers/quantum-error-correction/preview"3. Buy the full dossier with any x402 v2 client (TypeScript example):
import { wrapFetchWithPaymentFromConfig } from "@x402/fetch";
import { ExactEvmScheme } from "@x402/evm";
import { privateKeyToAccount } from "viem/accounts";
const account = privateKeyToAccount(process.env.AGENT_WALLET_KEY);
const fetchPaid = wrapFetchWithPaymentFromConfig(fetch, {
schemes: [{ network: "eip155:84532", client: new ExactEvmScheme(account) }],
});
const res = await fetchPaid("https://space-kitty.com/api/v1/dossiers/quantum-error-correction");
const dossier = await res.json(); // claims, timeline, benchmarks, bibliography…
const receipt = res.headers.get("X-Archive-Receipt"); // free re-downloads for 24hPrefer Python or another stack? Any x402-compatible client works. See payments for the wire format.
Endpoints
API index: endpoints, pricing, payment terms, archive statistics.
Research domains with topic and paper counts.
All topics (filter with ?domain=).
Parameters: domain
Topic overview: description, dossier preview, top and newest papers.
Ranked paper metadata for a topic, with abstracts where licensed.
Parameters: sort=top|recent, limit (≤100), offset, kind
Full-text search over papers, trials and dossiers with highlighted snippets.
Parameters: q, type=all|papers|dossiers, topic, kind, year_from, year_to, limit (≤50), offset
One paper or trial record (ids like arxiv:2408.13687, doi:10.1038/…, nct:NCT05000000).
Newest additions across all topics.
Parameters: topic, limit (≤50)
Dossier catalog with summaries, statistics and prices.
Parameters: domain
Free dossier preview: summary, contents, statistics, verification record, sample claims.
Full dossier: Markdown body, structured claims with citations, timeline, benchmarks, open problems, glossary, resolved bibliography. Send Accept: text/markdown for Markdown.
Research pack: the topic dossier plus the 50 most influential and 25 newest papers (with licensed abstracts) and related trials, in one JSON document.
Model Context Protocol endpoint (Streamable HTTP, JSON responses). Free search tools; paid tools settle via x402.
Formats & content negotiation
- Every HTML page also answers as Markdown: send
Accept: text/markdownor add?format=md. - API responses are JSON. The paid dossier endpoint also returns Markdown with
Accept: text/markdown. - Machine-readable descriptions: /llms.txt, /llms-full.txt, /openapi.json, /.well-known/x402.
- Paper ids:
arxiv:<id>,doi:<doi>,pmc:<PMCID>,nct:<NCT id>,openalex:<W id>.
Errors & limits
402: payment required or payment invalid. The body explains why and carries v1-compatible terms; v2 terms are in thePAYMENT-REQUIREDheader.404on a paid route: the resource does not exist and nothing was charged.429: slow down. Free endpoints are rate limited per client; paid requests are never rate limited after payment.503: payments temporarily unavailable. Nothing is charged.
You are never charged when the request fails: settlement happens only after the response is ready, and only if it succeeded.
Attribution & responsible use
Metadata from arXiv, OpenAlex, Europe PMC and ClinicalTrials.gov keeps its original license; each record says which. Dossiers are original works licensed to the purchaser for research use, including as context for AI systems, with attribution to Space Kitty Archive. Do not republish full dossiers. Medical content is not medical advice.