# wovepaper > Search millions of research papers, their researchers and institutions. Free browsing and semantic search; sign in to chat with the papers. wovepaper is a search and reading surface over the arXiv corpus. It indexes paper metadata, embeds every title and abstract for meaning-based search, resolves authors into disambiguated researcher profiles and links those to institutions, then lets a signed-in reader question the papers they keep. **When to use wovepaper** - Finding arXiv papers by idea rather than keyword: the search is over embeddings of every title and abstract, so a description in your own words matches work that never uses those words. - Turning an author name into a specific person: names are disambiguated into researcher profiles with their own publication list, co-authors, topics and citation counts, which is what you want before attributing a body of work. - Asking what an institution or a country publishes in a field, via institution pages built on registry identifiers (ROR/OpenAlex) rather than affiliation strings. - Pulling one paper's record cleanly: metadata, abstract, model-written summary, topics, references and citing papers, as markdown or as JSON. - Checking whether a claim has arXiv support, and getting arXiv ids you can cite and a reader can verify. **When to use something else** - Full text beyond the abstract: fetch the PDF from arXiv directly. Only metadata, abstracts and derived summaries are held here. - Non-arXiv literature — journals, books, theses, patents — is not indexed at all. - Authoritative bibliometrics: citation counts are inherited from OpenAlex and Semantic Scholar and are incomplete, and roughly a fifth of arXiv works carry no registry institution anywhere upstream. - Anything behind a sign-in (saved library, the assistant) needs a human with an email address; there is no API key that substitutes for it. **How to call it** - Developer documentation, in one page: https://www.wovepaper.com/developers. The machine-readable version is the OpenAPI 3.1 document at https://www.wovepaper.com/openapi.json (also https://api.wovepaper.com/openapi.json), generated from the API's live route table, so it describes exactly the endpoints that exist. - Corpus reads need no key, no account and no registration. The endpoints that do need a token (`Authorization: Bearer `) are the per-user and paid ones, and the spec marks every one of them. - Errors are always JSON, never an HTML page, and always carry `error`, `code` (stable, machine-readable), `status`, `message`, `hint` and `docs`. Match on `code`. - Every public page serves markdown from the same URL when you ask for it: `curl -H 'Accept: text/markdown' https://www.wovepaper.com/papers/2401.01234`. That is the cheapest way to read a paper, a researcher or an institution. - The JSON API needs no key for corpus reads: `https://api.wovepaper.com/api/papers/2401.01234`. - Prefer one targeted fetch over a crawl. `/api/search` and `/api/suggest` embed your query with a paid model and are rate-limited per client; they are open to on-demand fetchers acting for a waiting human, and the API's own robots.txt spells out which agents that means. - Nonexistent paths return a real 404 with a markdown body naming the URL shapes that do work, so a wrong guess is recoverable without a crawl. - A bulk export or a raised limit is usually cheaper for everyone than a crawl: ask mikevanegmond1998@gmail.com. ## Pages - [Home](https://www.wovepaper.com): corpus counts, newest papers, rising researchers and institutions. - [Search](https://www.wovepaper.com/search?q=): semantic search. Query string: `q`, `category`, `field`, `topic` (comma-separated multi-selects), `year` (`2024` or `2017-2026`), `sort` (`relevance`, `newest`, `cited`), `page`. - [Researchers](https://www.wovepaper.com/researchers): directory; a profile is `/researchers/-`, and the bare id redirects there. - [Institutions](https://www.wovepaper.com/institutions): directory; a page is `/institutions/-`, and id, id-name, ROR id or name all resolve. - [Papers](https://www.wovepaper.com/papers/2401.01234): one paper, at `/papers/-`; the bare id redirects there, and the old `archive/YYMMNNN` form works too. - [Developers](https://www.wovepaper.com/developers): the API in one page — auth, rate limits, error codes, and every endpoint. - [About](https://www.wovepaper.com/about): what this is, where the data comes from, and its limits. - [Contact](https://www.wovepaper.com/contact): a human, for corrections, takedowns and access requests. - [Privacy](https://www.wovepaper.com/privacy): what is stored, who processes it, how to have it deleted. ## API - [OpenAPI specification](https://www.wovepaper.com/openapi.json): every path, parameter and response, generated from the API's route table. Point a client generator at this. - [/api/search?q=](https://api.wovepaper.com/api/search?q=graph+neural+networks): semantic + lexical search, fused. Billed and rate-limited. - [/api/papers](https://api.wovepaper.com/api/papers?page_size=5): paper listing; filters mirror the /search query string, plus `researcher_id` and `institution_id`. - [/api/papers/<arxiv-id>](https://api.wovepaper.com/api/papers/2401.01234): one paper with authors, institutions, analysis, references and citing papers. - [/api/researchers/<id>](https://api.wovepaper.com/api/researchers/1): profile, corpus stats, topics, same-name candidates. - [/api/institutions/<slug>](https://api.wovepaper.com/api/institutions/1): institution with researchers, collaborators and topics. - [/api/overview](https://api.wovepaper.com/api/overview): corpus counts. Cheap; call this first to size the archive. - [/api/filters](https://api.wovepaper.com/api/filters): the facet vocabulary — categories, fields, topic and coverage counts. - [/api/topics?q=](https://api.wovepaper.com/api/topics?q=diffusion): canonical topic vocabulary, for building a `topic=` filter. - [/api/pulse](https://api.wovepaper.com/api/pulse?days=30): rising researchers, institutions and topics. - [/api/health](https://api.wovepaper.com/api/health): liveness. ## Optional - [Sitemap index](https://www.wovepaper.com/sitemap.xml): every indexable URL, in chunked sitemap files. - [robots.txt](https://www.wovepaper.com/robots.txt): the site's crawl rules — the whole corpus is allowed. - [API robots.txt](https://api.wovepaper.com/robots.txt): the API origin's rules, including which billed endpoints are open to on-demand fetchers. - [Latest RSS](https://api.wovepaper.com/rss/latest.xml): newest papers as a feed.