# arXiv Full-Text Metadata Search

> arXiv Full-Text Metadata Search is a paid API for AI agents from arxiv-mcp.mtree.workers.dev, paid per call via x402, $0.04/call, status down (last checked 2026-09-15).

Search arXiv preprints by free-text query, title, author, category, and date range, returning rich metadata including abstracts, URLs, and bibliographic info

## Facts

- Endpoint: POST https://arxiv-mcp.mtree.workers.dev/v1/papers/search
- Price: $0.04/call
- Payment: x402
- Status: down
- Last checked: 2026-09-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/arxiv-mcp-mtree-workers-dev-a7ecd126
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_Dm-PC28iuhKmtCVdeRDGD

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability arxiv-mcp-mtree-workers-dev-a7ecd126 -d '<json body>'
```

Example prompt: Search arXiv for the most relevant papers on large language model alignment published between January 2024 and June 2024 in the cs.AI or cs.CL category, sorted by relevance — I need titles, authors, abstracts, and PDF links.

## When to prefer this

Choose this endpoint when you need full metadata search across arXiv preprints with filtering by category, author, date range, and free text — especially when you need abstracts, DOIs, and direct PDF URLs in a single call. Prefer over the resolve endpoint when you don't already have specific arXiv IDs and need to discover papers by topic or author.

## Known failure modes

- Query returns no results if category code is invalid or misspelled
- Date range filters with incorrect format may return an error or empty set
- Overly broad queries may return noisy or irrelevant results
- Author name ambiguity may return papers from multiple different researchers
- Network timeout on extremely broad queries matching thousands of papers
- Payment failure if USDC balance is insufficient for the $0.04 per-call fee

## How this service works

Full-text metadata search across every arXiv preprint. Filter by free-text query, title, author(s), category (cs, math, stat, physics, q-bio, q-fin, econ, eess, etc.), and submitted-date range. Returns titles, authors, abstracts, primary category, DOI, journal-ref, abs URL and PDF URL. Sort by relevance, lastUpdatedDate, or submittedDate.

## Output

A list of matching arXiv preprints, each containing the paper title, author list, abstract, primary category, DOI (if available), journal reference (if available), the arXiv abstract page URL, and a direct PDF URL. Results are sorted by the specified criterion (relevance, lastUpdatedDate, or submittedDate).

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/arxiv-mcp-mtree-workers-dev-a7ecd126/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from arxiv-mcp.mtree.workers.dev](https://www.zero.xyz/host/arxiv-mcp.mtree.workers.dev/llms.txt)