wikipy

Local Wikipedia and Wikidata from on-disk dumps. Search and browse without hitting Wikimedia servers.

English Wikipedia Search and browse the English dump Wikipédia en français Rechercher et parcourir le dump français Wikidata Search entities and browse structured data by QID

How to use

Pick a corpus above, or call the JSON API directly. All data is served from local dumps — no requests to Wikimedia.

Browser

Use the search bars on each language or Wikidata home page. Results link to article pages or entity profiles.

GET /en/?q=…
Full-text search in the English dump (HTML results).
GET /fr/?q=…
Full-text search in the French dump.
GET /en/{path} · GET /fr/{path}
Raw article HTML from the ZIM file, e.g. /en/Albert_Einstein.
GET /wd/?q=…&filter=…
Wikidata entity search. Open a QID directly, e.g. /wd/Q42.

Wikipedia API

JSON endpoints under /api/{lang}/ where lang is en or fr.

Full-text search

Searches article bodies using the ZIM full-text index. Best for topics and phrases.

GET /api/en/search?q=…
Returns matching articles with paths and links to HTML and Markdown.

Parameters: q (required) · limit (default 5, max 100)

curl -sG 'https://wiki.corbia.net/api/en/search' \
  --data-urlencode 'q=general relativity' \
  --data-urlencode 'limit=5'

Title suggestions

Prefix / title lookup using the ZIM title index. Best when you know roughly what the article is called.

GET /api/en/suggest?q=…
Returns title matches ranked by the suggestion index.

Parameters: q (required) · limit (default 5, max 100)

curl -sG 'https://wiki.corbia.net/api/en/suggest' \
  --data-urlencode 'q=Albert Einstein'

Markdown article

Fetches an article by exact ZIM path and returns clean Markdown converted from the local HTML.

GET /api/en/md/{path}
JSON with a markdown field plus metadata (title, wikipedia_url, chars).
curl -s 'https://wiki.corbia.net/api/en/md/Albert_Einstein' \
  | jq -r '.markdown' > Albert_Einstein.md

Wikidata API

JSON endpoints under /api/wd/, backed by a local SQLite database.

Entity search

GET /api/wd/search?q=…
FTS5 full-text search over entity names and descriptions.

Parameters:

curl -sG 'https://wiki.corbia.net/api/wd/search' \
  --data-urlencode 'q=macron' \
  --data-urlencode 'filter=person' \
  --data-urlencode 'limit=5'

Entity lookup

GET /api/wd/{qid} · GET /api/wd/{qid}/raw
Full Wikidata entity JSON (both routes return the same payload).
GET /api/wd/{qid}/summary
Readable profile: labels, description, instance-of, birth date, occupation, and other common claims.
curl -s 'https://wiki.corbia.net/api/wd/Q42/summary' | jq '{id, name, instance_of}'

Response formats

Route Format Use for
/en/{path} HTML Browser viewing, debugging
/api/{lang}/search · /api/{lang}/suggest JSON Finding articles by query
/api/{lang}/md/{path} JSON (markdown field) Agent-readable article text
/api/wd/search JSON Finding entities; add json=true for full blobs
/api/wd/{qid} JSON (Wikidata entity) Structured facts, claims, sitelinks
/api/wd/{qid}/summary JSON (profile) Quick human- or agent-readable overview

Typical workflow

  1. Search: /api/en/search?q=… or /api/wd/search?q=…
  2. Pick a result path or id
  3. Fetch content: /api/en/md/{path} for Wikipedia, /api/wd/{qid}/summary for Wikidata

URL-encode query strings in real requests. With curl, use --data-urlencode for q values that contain spaces or accents.