Academic paper search across arXiv, Semantic Scholar, and PubMed. Normalized JSON results with structured citations — no HTML parsing, no API key management, just clean data.
Your agent needs to find academic papers. Normally that means hitting arXiv's Atom XML API, Semantic Scholar's JSON API, and PubMed's two-step E-utilities — each with different schemas, different auth, different rate limits. PAPERS does it all in one call.
Free tier only. Parses Atom XML into structured JSON with title, authors, abstract, categories, and paper URL. No API key needed.
Paid tier. Rich metadata including citation counts, publication year, and open access badges. Rate-limited but fast.
Paid tier. NCBI's biomedical literature via E-utilities. Free tier uses esearch only; paid adds full efetch for complete metadata.
24-hour cache on every query. Same search requested 100 times = 1 upstream API call. SQLite-backed, no Redis needed.
All three sources are free to query (with rate limits). PAPERS handles authentication, parsing, and normalization — so your agent gets one consistent schema.
Open-access preprints. No API key. 200 queries/hour rate limit. Returns Atom XML parsed into JSON. Covers CS, physics, math, bio, and more.
AI-powered academic search. Free tier: 100/min. Paid: higher limits. Returns structured JSON with citation graphs, influence scores, and PDF links.
Biomedical literature database. Free tier: esearch only (titles/abstracts). Paid: adds efetch for full metadata. 3 requests/second limit.
Search arXiv only. Free tier: 10/day/IP, max 5 results.
curl "https://papers.4bots.ai/v1/free/papers/search?q=speculative+decoding&limit=5"
Example response:
{
"query": "speculative decoding",
"sources": ["arxiv"],
"results": [
{
"title": "Speculative Decoding: Experimental...",
"authors": ["Smith, J.", "Lee, K."],
"abstract": "We present a novel approach...",
"categories": ["cs.LG", "cs.CL"],
"published": "2024-03-15",
"arxiv_id": "2403.09876",
"url": "https://arxiv.org/abs/2403.09876"
}
],
"total_found": 42,
"cached": false,
"disclaimer": "Results for informational and research purposes only. Preprint results from arXiv have not been peer-reviewed.",
"free_remaining": 7
}
Search all three sources in parallel. Paid tier: $0.003/search, max 20 results.
curl "https://papers.4bots.ai/v1/paid/papers/search?q=speculative+decoding&limit=20"
Same schema. Three times the results.
Health check. Always free, always returns 200.
curl "https://papers.4bots.ai/health"
10 free searches per day covers development and light agent workloads. Paid tier unlocks Semantic Scholar + PubMed.
No signup. No API key. Just send a request.
How x402 works