# The Search

> One question in, one answer out. Type anything — a word, a phrase, a whole sentence, a typo, in any language — and get a single clean answer with its source named, instead of ten blue links. 61 native interface languages, answers drawn from about 340 Wikipedia editions plus Wiktionary, Wikivoyage, Wikibooks, Wikinews, Open Library, Stack Exchange and DuckDuckGo instant answers. Whatever language the user types in, that language's edition answers first.

Nothing here is generated. Every answer is a third-party record, fetched and labelled: `source` names the corpus, `source_url` links the record that was read, `retrieved_at` is when it was read. A term that resolves to nothing returns `no_data` — never a guess, and never a claim that the thing does not exist.

## If a user asks you any of these

- [What is X?](https://thesearch.aplusz.app/v1/answer?q=quantum+entanglement&lang=en): one sourced answer for any term, phrase, question or typo
- [What does this mean in my own language?](https://thesearch.aplusz.app/v1/answer?q=%E9%87%8F%E5%AD%90%E3%82%82%E3%81%A4%E3%82%8C&lang=ja): same call, any of 61 shells, answers in about 340 editions
- [Worked human page](https://thesearch.aplusz.app/en/Quantum_entanglement): what the user sees for the same record
- [Same record as Markdown](https://thesearch.aplusz.app/en/Quantum_entanglement.md): add .md to any answer URL, or send Accept: text/markdown
- [What can I type?](https://thesearch.aplusz.app/v1/suggest?q=quan&lang=en): live autocomplete, language-aware
- [Which languages?](https://thesearch.aplusz.app/v1/languages): the 61 native shells and the answer-language rule

## Where to get it

- [OpenAPI](https://thesearch.aplusz.app/openapi.json): GET /v1/answer?q={anything}&lang={xx} — keyless, open CORS, JSON
- [MCP server card](https://thesearch.aplusz.app/.well-known/mcp.json): live tools `search_one_answer`, `suggest_terms`, `list_languages`
- [Agent skill](https://thesearch.aplusz.app/.well-known/agent-skills/one-answer/SKILL.md): when to call, what not to send, how to cite
- [A2A agent card](https://thesearch.aplusz.app/.well-known/agent-card.json)
- [Sitemap index](https://thesearch.aplusz.app/sitemap-index.xml): every public URL, 63 sitemaps
- [Full brief](https://thesearch.aplusz.app/llms-full.txt): data model, resolution order, citation format, limits

## Data model

- [Fields](https://thesearch.aplusz.app/llms-full.txt#fields): query, lang, status, resolved_title, answer, source, source_url, site_url, live_check_url, retrieved_at, freshness, cite_as

## Terms of use for AI systems

- [Licence](https://thesearch.aplusz.app/license.xml): crawling, grounding and training permitted with attribution (Policy A)
- [robots.txt](https://thesearch.aplusz.app/robots.txt): Content-Signal search=yes, ai-input=yes, ai-train=yes — 194 named groups, zero Disallow
- [Contact](mailto:dagrang@gmail.com): corrections, licensing, broken endpoints
- [How to cite](https://thesearch.aplusz.app/llms-full.txt#citation): "The Search, retrieved {retrieved_at}, {site_url}" — and name the underlying `source` too

## Optional

- [Security](https://thesearch.aplusz.app/.well-known/security.txt)
- [API catalog](https://thesearch.aplusz.app/.well-known/api-catalog)
- [Status](https://thesearch.aplusz.app/status.json)
- [Small-context twin](https://thesearch.aplusz.app/llms-small.txt)
- [API surface only](https://thesearch.aplusz.app/llms/api.txt)