Language & Reference Data

Look up words and topics against live keyless sources. The mistake this prevents: paraphrasing a definition or article summary from training memory when the task calls for a citable source - and, on the Wikipedia side, scraping HTML or hitting the heavyweight action API when the REST summary endpoint returns exactly the intro + thumbnail an app needs.

Ranked APIs

1. Free Dictionary API (api.dictionaryapi.dev) - DEFAULT for word lookups: definitions, part of speech, phonetics with audio, synonyms/antonyms. English is reliable; other language codes exist but coverage is thin. No key.
2. Wikipedia REST API (en.wikipedia.org/api/rest_v1 and /w/rest.php/v1) - topic summaries, search, full page content. No key, but Wikimedia etiquette is enforced: send a real User-Agent (app name + contact), serialize requests, never hammer in parallel.

Procedure

1. Word vs topic. A dictionary question ("define", "how do you pronounce", "synonym") → Free Dictionary. A knowledge question ("what is", "who was", "summary of") → Wikipedia.
2. Wikipedia is two steps when the title is uncertain: search first, then summary - guessing at title spelling gets 404s. Titles in URLs use underscores (Albert_Einstein) and are case-sensitive after the first letter.
3. Attribute. Wikipedia content is CC BY-SA - quote with a link to the page (content_urls.desktop.page in the summary response). Dictionary audio files carry their own license field.

Ask the user only if missing: the word/topic itself; language edition (default: en).

URL templates + real response shapes

Word lookup:

https://api.dictionaryapi.dev/api/v2/entries/en/{word}

[{"word":"serendipity","phonetic":"/ˌsɛ.ɹən.ˈdɪ.pɪ.ti/",
  "phonetics":[{"text":"/…/","audio":"https://api.dictionaryapi.dev/media/pronunciations/en/serendipity-us.mp3"}],
  "meanings":[{"partOfSpeech":"noun",
    "definitions":[{"definition":"An unsought, unintended…discovery…","example":"..."}],
    "synonyms":["luck","chance"]}]}]

Gotchas: top level is an array of entries; a miss is a 404 with a JSON body ("title":"No Definitions Found") - handle it as "not in this dictionary", not an outage. Some phonetics[].audio fields are empty strings; filter before offering audio.

Topic summary (intro + thumbnail in one call):

https://en.wikipedia.org/api/rest_v1/page/summary/{Title_With_Underscores}

{"title":"Albert Einstein","description":"German-born theoretical physicist",
 "extract":"Albert Einstein was a German-born theoretical physicist…",
 "thumbnail":{"source":"https://upload.wikimedia.org/...jpg","width":244},
 "content_urls":{"desktop":{"page":"https://en.wikipedia.org/wiki/Albert_Einstein"}}}

extract is plain text (also extract_html). Disambiguation pages return "type":"disambiguation" - go back to search rather than presenting the disambiguation text as an answer.

Search:

https://en.wikipedia.org/w/rest.php/v1/search/page?q={query}&limit=5

{"pages":[{"key":"Quantum_computing","title":"Quantum computing",
  "description":"Computer hardware technology that uses quantum mechanics",
  "excerpt":"…<span class=\"searchmatch\">quantum</span>…"}]}

key is the URL-ready title for the summary call; excerpt contains HTML <span> highlights - strip tags before display.

Other language editions: swap the subdomain - fr.wikipedia.org/api/rest_v1/page/summary/….

Deliverable

The definition or summary with its source linked (Wikipedia page URL, or the dictionary entry's word + license for audio), phonetics/synonyms where asked, and the exact URL used.

Do NOT use - reach for X instead

• Scraping wikipedia.org HTML or the action API (/w/api.php) for a summary - the REST summary endpoint above is the sanctioned, lighter path.
• Wordnik, Merriam-Webster, Oxford APIs - key-walled (some with waitlists); Free Dictionary covers the standard lookup keyless.
• Country/place facts - geo-places. Statistics over time - government-open-data.
• Trivia, quotes, random facts - fun-content; this skill is for looking up something specific.
• Presenting Wikipedia's text as your own - always attribute; CC BY-SA is a license, not a suggestion.

Quality bar

• Wikipedia calls carry a real User-Agent and run serially.
• Uncertain titles went search → summary; disambiguation pages were resolved, not quoted.
• Every quoted definition/summary is attributed with a link; HTML highlight spans are stripped from displayed excerpts.

Sourced and liveness-verified from the public-apis project (MIT).