Wikidata Lexemes Scraper

Wikidata Lexemes Scraper

Scrape Wikidata lexemes by lemma, language, or lexical category, up to a million per run. Each lexeme returns its ID, lemma, language, lexical category, senses, and optional inflected forms. No API key or complex SPARQL queries. Export to CSV, JSON, Excel, or XML.

Wikidata's lexicographical data is locked behind SPARQL endpoints and the MediaWiki API, which require query expertise and pagination handling. This Actor reads lexemes directly from the public Wikidata search index, filtered by lemma string, language QID, or lexical category QID, and returns each match in one flat, consistent schema. It is the fastest way to build a multilingual dictionary dataset without writing a single query.

Who uses it	What they scrape Wikidata for
Computational linguists	Build a structured lexicon for a specific language to train NLP models.
Language learning app developers	Gather vocabulary lists with definitions and grammatical categories for a new course module.
Lexicographers	Extract a complete set of lemmas and their senses for a dictionary project.
Data journalists	Analyze the coverage and gaps of lexicographical data across different languages on Wikidata.

What it does

This Actor collects Wikidata lexemes by lemma search, language, or lexical category, and returns each one as a flat row with its senses and optional inflected forms.

• 🔍 Lemma search: Find lexemes by their base form, like 'run', 'maison', or 'こんにちは', across all languages.
• 🌐 Language filter: Restrict results to a single language using its Wikidata QID, such as Q1860 for English or Q150 for French.
• 🏷️ Lexical category filter: Narrow results to a part of speech, like Q1084 for nouns or Q24905 for verbs.
• 📋 Optional inflected forms: Include the full list of inflected forms (plurals, conjugations, declensions) on each lexeme record.

Results export to CSV, JSON, Excel, or XML, or straight from the API.

What you can do with Wikidata data

📚 Build a monolingual dictionary.

A lexicographer searches for all English lexemes, filters by noun and verb categories, and exports the dataset to seed a new online dictionary.

🤖 Train a multilingual NLP model.

A computational linguist collects lexemes for French, Spanish, and German with their inflected forms to create a training corpus for a lemmatizer.

📊 Audit Wikidata's language coverage.

A data journalist scrapes lexemes for several underrepresented languages and compares the counts to assess data completeness on Wikidata.

🎓 Create language learning flashcards.

An app developer gathers the 1,000 most common Japanese lexemes with their senses and forms to populate a vocabulary drill module.

Why choose this scraper

	What you get
No SPARQL required	Get lexicographical data without writing or debugging complex queries.
Flat, consistent schema	Every lexeme, sense, and form is returned as a predictable row, ready for analysis.
Multilingual by default	Search across all languages or restrict to one, with native script support.
Scales to a million lexemes	Collect small samples or the entire lexeme namespace in a single run.

How it compares

No other Store actor targets Wikidata the same way, so the honest comparison is with the alternatives teams actually weigh.

	Wikidata Lexemes Scraper	Build it in-house	By hand
Setup	Run it now, zero config	Days of engineering	None, but hours per pull
When Wikidata changes	Maintained for you	You fix it	You re-learn the page
Proxies, retries, anti-bot	Built in	Your problem	Browser only
Output	Fixed JSON schema, CSV/Excel export	Whatever you build	Copy-paste
Cost	Pay per result	Engineering time	Analyst hours

Configure the run

Drive the Actor from a lemma search query, and optionally post-filter results by language QID or lexical category QID so only matching lexemes reach your dataset. The Input tab lists every parameter.

A first run with the defaults:

{
  "maxItems": 10,
  "searchQuery": "run",
  "searchLanguage": "en"
}

A larger pull:

{
  "maxItems": 200,
  "searchQuery": "run",
  "searchLanguage": "en"
}

Pricing

Pay-per-result: $0.01333 per result collected. You pay only for the results written to your dataset.

Results collected	Approximate cost
100 results	$1.33
1,000 results	$13.33
10,000 results	$133.30

New Apify accounts start with $5 in free credit.

Free users

Free-plan runs return up to 10 results as a preview. Upgrade your Apify plan to collect up to 1,000,000 results per run.

Run it

1. Create a free Apify account with $5 in credit.
2. Open the Wikidata Lexemes Scraper.
3. Set your inputs and any filters, then click Start.
4. Export the results as CSV, Excel, JSON, or XML from the Dataset tab.

Run it programmatically through the Apify API (run-sync-get-dataset-items) or the ApifyClient for JavaScript and Python.

Use with AI agents (MCP)

Give an AI agent live access to Wikidata through the Model Context Protocol. Add the Actor to Claude, Cursor, or any MCP client:

$claude mcp add --transport http apify "https://mcp.apify.com?tools=parseforge/wikidata-lexemes-scraper"

Then prompt it in plain language to run the scraper and read back the results.

Troubleshooting

Why am I getting no results?

Check your search query spelling and try a simpler lemma. If you are using language or lexical category filters, verify the QIDs are correct on Wikidata.org. Remove the filters to test if the search itself returns results.

Why are my results in the wrong language?

The search index may return lexemes in multiple languages for the same lemma string. Use the language QID filter to restrict results to a single language.

The run stopped before reaching my maxItems limit.

The Actor stops when the Wikidata search index has no more matching lexemes. Try a broader search query or remove the language and lexical category filters to get more results.

Why are inflected forms missing from my output?

Make sure the 'Include Forms' checkbox is enabled in the input. By default, forms are not embedded to keep the records smaller and faster to collect.

I got a timeout error.

The Wikidata search API may be slow for very large result sets. Try reducing the maxItems limit and running the Actor in smaller batches.

FAQ

Question	Answer
What is a Wikidata lexeme?	A lexeme is a unit of lexical meaning in Wikidata, representing a word or phrase. It contains a lemma, a language, a lexical category, senses (definitions), and optionally inflected forms like plurals or past tense.
Do I need a Wikidata API key?	No. This Actor reads from the public Wikidata search index and requires no authentication, API key, or app registration.
How do I find the QID for a language or lexical category?	Search for the item on Wikidata.org. The QID is in the page URL and title. For example, English is Q1860, noun is Q1084, and verb is Q24905.
Can I search for lexemes in non-Latin scripts?	Yes. The search query accepts any Unicode string, so you can search for lemmas in Arabic, Chinese, Cyrillic, or any other script.
What are inflected forms?	Inflected forms are the grammatical variants of a lexeme, such as plural nouns, conjugated verbs, or declined adjectives. Enable the 'Include Forms' option to embed them in each lexeme record.
How many lexemes can I scrape in one run?	You can set the maximum from 1 up to 1,000,000 lexemes per run. The Actor stops when it reaches your limit or exhausts the search results.
Does this Actor use the Wikidata SPARQL endpoint?	No. It uses the MediaWiki search API for the Lexeme namespace, which is faster for lemma-based searches and avoids SPARQL timeout issues.
Can I get lexemes for multiple languages at once?	The search query runs across all languages by default. To restrict to one language, use the language QID filter. For multiple languages, run the Actor separately for each QID.
What output formats are supported?	You can export your dataset to CSV, JSON, Excel, or XML from the Apify dataset tab.
How do I get only nouns or only verbs?	Enter the lexical category QID in the optional filter field. For example, Q1084 for nouns or Q24905 for verbs.

Related actors

Browse the full ParseForge collection for more scrapers.

🆘 Need help? Email parseforge@protonmail.com with your run ID, your input, and what you expected.

⚠️ Disclaimer. This Actor is unofficial and is not affiliated with, endorsed by, or sponsored by Wikimedia Foundation, Inc. It collects only publicly available data. You are responsible for using the collected data in compliance with the source's terms of service and applicable data-protection laws, including GDPR, CCPA, and PIPL. Do not use it to collect personal data unlawfully.