Editorial Policy

Purpose and ownership

Word Helper uses source-tracked lexical fields and structured word information to create practical reference pages and tools. The website is created and maintained by Jay Sudha (Shatanjay Sudha), a builder of practical online tools and educational utilities. The editorial goal is to make word information easier to understand, compare, and use for students, writers, learners, and word-game players. The site does not claim to be an official dictionary or an academic authority.

How word content is created

Word Helper separates content by provenance. Eligible static lexical fields may come from the Datamuse API, a composite service whose upstream sources and terms vary by field. The Free Dictionary API is used only for best-effort runtime lookup until the provenance and licensing of its returned fields are verified for static indexable reuse. Word Helper's own guides, tool explanations, curated notes, and generated candidates are identified separately rather than presented as third-party source text.

Word Helper's editorial work includes selection, standardization, screening, and structuring — choosing which eligible words and senses to surface, formatting records consistently, filtering malformed or low-quality results, and applying the quality gate below. Generated example candidates are automatically screened, but that is not represented as human review. We do not claim to have independently authored sourced definitions or to have verified one blanket license for every field returned by a composite service.

How Word Helper builds word pages

Word Helper publishes a large library of word pages. Here is honestly how they are built, so readers and search engines understand what they are:

• Word data is structured from eligible, source-tracked lexical fields (see above) and formatted into one consistent template for practical, fast lookup.
• A page may include a definition, a plain-English "in simple terms" line, example sentences, synonyms, antonyms, related words, rhymes, syllables, part of speech, and learning notes — but not every word has every section. Sections appear only when reliable data exists for them.
• Where data is missing or uncertain, the page uses a graceful fallback rather than an empty box, and invites a correction.
• Priority word pages — common, high-intent words — receive additional editorial formatting: a closer look at meaning, how close synonyms differ in tone, common confusions, and examples at different levels. This richer layer is added by hand for a curated set of words and grows over time.
• Every page carries a source and correction note, and users can report an issue at any time.
• Word Helper does not claim official dictionary status; it is an educational reference that structures source-tracked language data and its own editorial material for easier use.

How data sources are handled

Word Helper uses the verified public-domain ENABLE word list as the production word-inventory backbone for eligible local tools. The first 172,823 lines of the repository inventory match the recorded ENABLE source exactly; membership supplies a word string, not a definition, editorial endorsement, or game-specific acceptance claim. The later supplementary system-list segment has unverified origin and redistribution terms and is excluded from production. Rhyme and syllable tools use clearly labelled pattern estimates rather than source-list membership as proof of pronunciation.

How corrections are reviewed

Content corrections can be submitted by email to hello@wordhelper.online. Reports of factual errors, definition inaccuracies, broken tool results, or misleading content are treated as high priority. Corrections are reviewed, verified, and applied to the source files. Updated pages are rebuilt and redeployed as quickly as possible. We do not wait for a scheduled update cycle to fix confirmed errors.

Sources, licensing, and attribution

Word Helper records each source separately and limits it to uses supported by its provenance:

• Production word inventory: the first 172,823 lines of src/data/words.txt exactly match the verified public-domain ENABLE list. Eligible Word Helper seed and published entries may also be included. The later supplementary system-list segment is not production-eligible and is excluded.
• Lexical fields and relations: the Datamuse API is a composite service drawing on sources including WordNet, Wiktionary, CMUdict, and corpus data. Source and licensing requirements vary by field, so static reuse is eligible only where field-level provenance and required attribution are recorded; no blanket license claim is applied.
• Runtime dictionary lookup: the Free Dictionary API may provide a best-effort live result. Its returned-field provenance and licensing have not been verified for static indexable reuse, so it is not used for that purpose.
• Frequency ordering: the Norvig count_1w file is used only as a non-visible ranking signal, not as a word-validity source, definition source, or current usage statistic.
• Word Helper material: original guides, tool explanations, curated lists, editorial notes, corrections, and generated candidates are maintained separately from third-party lexical fields. Generated candidates are never presented as quoted source text or blanket human-reviewed content.

Word Helper does not scrape or republish content from commercial dictionaries. If you believe any entry overlaps with material you hold rights to, contact us and we will review and revise or remove it promptly.

Quality standards and indexation stages

Every generated word entry is classified by the same automated Tier A/B/C gate used by the build, sitemap, internal links, and audit reports:

• Tier A — Stage 1 indexable: a valid sourced definition, real part of speech, pronunciation and syllables, at least two natural examples that use the headword or a valid form, useful cleaned relations, complete metadata, and a genuinely curated or word-specific editorial layer. Only Tier A pages are linked from public A–Z hubs and included in XML sitemaps.
• Tier B — useful but not index-ready: a usable source entry that lacks one or more Tier A requirements. Tier B pages use noindex, follow, are excluded from sitemaps and public browse links, and can be promoted only after their data and editorial value improve.
• Tier C — failed or invalid: an invalid headword, placeholder, weak or missing definition, raw markup/array artifact, truncated definition, unsupported cross-reference, or missing source attribution. Tier C URLs return a real 404 and are not presented as word profiles.

Stage 2 may later admit strong specialist and advanced Tier B words after validation. Stage 3 is reserved for the remaining corpus after automated checks and sampled manual review. The tool dictionary, available word-profile records, index-eligible profiles, and editorially enriched profiles are deliberately separate counts.

How AI tools are used in content production

Word Helper uses AI language models as an aid in the content process in two specific, limited ways:

• Example-sentence generation: When a word page lacks a human-authored example sentence from the source APIs, an AI language model is used to generate a candidate sentence. All AI-generated sentences are then automatically screened against a clarity and accuracy checklist before appearing on the site.
• Standardization assistance: AI tools assist with normalizing data formats, checking for obvious inconsistencies, and flagging entries that need human review.

AI is not used to author definitions, etymologies, or content represented as third-party source text. Eligible sourced definitions retain field-level provenance; Free Dictionary results remain runtime-only under the current evidence state. If you spot an example sentence that reads as unnatural or inaccurate, please report it and it will be reviewed and corrected or replaced.

FAQ and schema accuracy

Every FAQ visible on a Word Helper page is reflected accurately in the page's FAQ schema. Schema is not added for content that is not visible to the user. Tool explanations in schemas reflect the actual logic used by the tool on that page.

How to report a content issue

Send a description of the issue to hello@wordhelper.online, including the URL of the page and as much detail as possible about the error. All reports are read and responded to within a few business days.