Words API
Utilities & PersonalLook up English words through Words API, retrieving definitions, synonyms, antonyms, examples, rhymes, syllables, and pronunciation. Agents check and vary language before content is published.
What This Integration Enables
A language model will produce a synonym, a rhyme or an etymology on request, and it will produce one whether or not the relation exists. That is fine for a first draft and not fine for anything a reader will trust. Words API is the reference an agent consults instead of guessing: definitions grouped by part of speech, synonyms and antonyms, example sentences, IPA pronunciation, syllable breakdowns, rhymes, usage-frequency scores, and the full set of WordNet lexical relations from broader and narrower terms through to component parts and regional usage. Giving an agent a lookup it can be held to is a different discipline from giving it a better prompt.
The relations are the part people underestimate. Broader and narrower terms turn a single tag into a category hierarchy. Get Regions Of Use tells you that a word is British before your American reader tells you. Get Usage Of tells you that a term is a trademark rather than a generic noun, which is the difference between copy and a letter from a legal department. Get Frequency returns a Zipf score you can grade vocabulary against, which is how a reading-level target stops being a matter of opinion. Search Words matches on spelling patterns, pronunciation patterns, part of speech and frequency range together, so a puzzle, a vocabulary list or a naming shortlist becomes one query rather than a scraping project.
Without FlowRunner
With FlowRunner
Use Case Scenarios
An editorial pass that shows its working
A content team runs every draft through a repetition check before publication. An agent reads the draft from Contentful, identifies terms used more often than the house style allows, and calls Get Definitions and Get Synonyms for each. Rather than rewriting the sentence, it assembles a proposal: the original word, the sense it carries in this sentence, and two or three candidate replacements with their own definitions attached. The editor sees the reasoning and makes the call. The agent applies the accepted swaps back to the draft and leaves the rejected ones alone.
Grading vocabulary against a reading level
An education publisher needs learner material to stay inside a target reading level. An agent walks each paragraph, calls Get Frequency for the content words and Get Syllables for the long ones, and scores the passage on the Zipf values and syllable counts it gets back. Where a passage exceeds the target, the agent uses Get Synonyms filtered by frequency to propose simpler alternatives, and Get Examples to check that the simpler word still reads naturally in a comparable sentence. The score comes from a dictionary rather than from a formula somebody invented in a spreadsheet.
Localizing copy between English markets
A product launch written for the United Kingdom is being adapted for the United States. An agent runs the copy through Get Regions Of Use and Get Usage Of, which surface the words the dictionary records as regional, slang, colloquial or trademarked. It does not silently substitute them. It builds a list, and where a regional term has a documented counterpart it pairs the two using Get Regional Words. The localization owner reviews the list, approves the replacements that hold and rewrites the ones that do not, and only then does the agent push the adapted copy to WordPress.
Human-in-Loop Highlight
Get Synonyms returns a flat list covering every sense of the word. That is not a defect, it is how a thesaurus works, but it means the API cannot tell an agent which meaning a given synonym belongs to, and a word with three senses will return substitutes for all three mixed together. Swap the wrong one into published copy and the sentence still parses; it just says something else. The related failure is quieter and more expensive: a candidate that Get Usage Of marks as a trademark, dropped into marketing copy as though it were a generic noun. So the agent never applies a substitution on its own authority. It stops with the proposal assembled, sends it to the content owner in Slack, and asks: "Nine repeated terms, 21 candidate replacements. Two are marked as trademarks and four are recorded as British usage. Each candidate is listed with the definition it came from. Approve the swaps you want and I will apply only those." Reference connectors like this one exist to make the agent's suggestion checkable, and human-in-the-loop is what keeps the check in front of the publish rather than behind it.
Agent Capabilities
32 actionsWord Lookup
4- Get Word Retrieves everything the dictionary holds for a single English word in one call: every definition grouped by part of speech, the synonyms and lexical relations attached to each definition, the syllable breakdown, the IPA pronunciation and the overall usage-frequency score. Use it when you want the complete record rather than one specific detail type.
- Get Word Details Retrieves one specific detail type for a word, chosen from the full list of detail types including definitions, synonyms, antonyms, examples and every lexical relation. Use it as a single flexible call when the detail type is decided at runtime rather than at design time.
- Get Definitions Retrieves every dictionary definition of a word, each returned with the part of speech it applies to. A word with several senses returns one entry per sense, which makes this the fastest way to disambiguate a term before an agent uses it anywhere else in a flow.
- Get Examples Retrieves example sentences showing the word used in natural context. Useful for language-learning content, for grounding a prompt, or for verifying that a chosen synonym still reads correctly in a sentence. Not every word in the dictionary carries examples.
Thesaurus
4- Get Synonyms Retrieves words that can be interchanged with the original word in the same context. The list is flat across all senses of the word, so an agent should read the definitions first whenever a term carries more than one meaning.
- Get Antonyms Retrieves words carrying the opposite meaning of the original word. Antonyms are recorded for a limited part of the dictionary, mostly adjectives and verbs, so an empty result is a normal outcome rather than a failure.
- Get Similar Words Retrieves words close in meaning to the original without being direct synonyms, the relation WordNet calls similarTo. It is most commonly populated for adjectives and is useful when a substitution should shift the shade of a word rather than replace it outright.
- Get Related Phrases Retrieves multi-word phrases that the original word belongs to, the relation WordNet calls also. It is a good source of phrasal verbs and idiomatic constructions, which are exactly the constructions a non-native draft tends to get slightly wrong.
Pronunciation And Sound
3- Get Pronunciation Retrieves the International Phonetic Alphabet transcription of a word. The result is keyed by part of speech when pronunciation differs between roles, and uses a single key when the word sounds the same in every case. IPA stress marks use a single quote character, which some parsers need to escape.
- Get Syllables Retrieves the syllable breakdown of a word, returning the count together with the ordered list of syllable fragments. Useful for hyphenation, for poetry and lyric work, and for readability scoring where syllable count feeds the formula.
- Get Rhymes Retrieves words that rhyme with the given word. Rhymes are never included in a normal word lookup, so this is the only way to obtain them. Results are grouped by part of speech, so a word pronounced differently as a noun and a verb returns separate groups.
Usage Statistics
1- Get Frequency Retrieves usage-frequency statistics for a word: the Zipf score on a log scale where higher means more common, the expected occurrences per million words of English text, and a diversity value for how likely the word is to appear in any given document. Use it to grade vocabulary difficulty or to filter rare terms out of a candidate list.
Lexical Relationships
12- Get Broader Terms Retrieves the more generic words the original word is a kind of, known as hypernyms. Follow the relation upwards to generalize a term or to build the upper levels of a category hierarchy.
- Get Narrower Terms Retrieves the more specific words that are kinds of the original word, known as hyponyms. Follow the relation downwards to expand a category into concrete examples, which is how a single tag becomes a usable taxonomy branch.
- Get Containing Wholes Retrieves the larger wholes the original word forms part of, known as holonyms. Use it to move from a component to the assemblies that contain it.
- Get Component Parts Retrieves the components that make up the original word, known as meronyms. Use it to decompose an object into its constituent pieces, for example when generating a parts checklist from a product noun.
- Get Instance Of Retrieves the general classes the original word is a named example of. The relation applies mainly to proper nouns such as people, places and organizations, which makes it useful for classifying entities pulled out of free text.
- Get Instances Retrieves the named examples of the original word. Use it to turn a general role or class into a list of real-world individuals.
- Get Member Of Retrieves the groups the original word belongs to. The relation is commonly used for biological taxonomy and for organizational membership.
- Get Members Retrieves the individual members of the group named by the original word. Use it to expand a collective noun into the entities it contains.
- Get Substance Of Retrieves the larger substances the original word is a constituent of. Use it to move from a material to the mixtures it appears in.
- Get Substances Retrieves the substances the original word is made of. Use it to break a material down into its chemical or physical constituents.
- Get Entailments Retrieves the actions implied by the original word. The relation applies almost exclusively to verbs and is useful when an agent needs to reason about what must also be true when an action happens.
- Get Pertains To Retrieves the words the original word is relevant to. The relation typically links a relational adjective back to the noun it derives from.
Domains And Regions
6- Get Parent Category Retrieves the domain category the original word belongs to. Use it to detect which subject field a specialized term comes from before deciding whether the audience will recognize it.
- Get Subcategories Retrieves the categories that fall under the original word. Use it to enumerate the sub-fields of a subject domain, for example when planning coverage across a topic.
- Get Usage Of Retrieves the usage domains the original word is an example of. This is how an agent detects that a term is slang, a trademark, a colloquialism or another marked register, which is the check that belongs in front of any published substitution.
- Get Domain Usages Retrieves the words belonging to the usage domain defined by the original word. Use it to harvest slang, idioms or trademarks as a set, for example when building a filter list.
- Get Regions Of Use Retrieves the geographic regions where the original word is used. Use it to flag regional vocabulary before copy is localized for another English-speaking market.
- Get Regional Words Retrieves the words associated with the region named by the original word. Use it to build a vocabulary list for a specific English-speaking locale, or to find the counterpart of a term flagged as regional.
Word Search
2- Search Words Searches the dictionary on any combination of spelling pattern, pronunciation pattern, part of speech, available detail types and usage-frequency range. It returns a paged list of matches with the total count, which is what makes word games, crossword helpers and vocabulary lists a single query rather than a project.
- Get Random Word Returns one random word with its full record, including definitions, syllables, pronunciation and frequency. Every Search Words filter is supported, so a draw can be biased towards words of a given length, sound, part of speech or frequency. Leave the filters empty to draw from the whole dictionary.
Frequently Asked Questions
What can FlowRunner do with Words API?
FlowRunner agents can run Get Word, Get Word Details, and Get Definitions in Words API, plus 29 more actions.
Does connecting Words API to FlowRunner require OAuth?
No. Words API connects to FlowRunner with an API key, no OAuth flow required.
Can Words API trigger a FlowRunner workflow automatically?
Words API doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with Words API
$100 in credits. No card required. Connect in minutes.