JokeAPI

• Get a jokeGet one joke, filtered however you need: by category, by language, one-line or two-part, with flagged jokes excluded and safe mode on. The joke comes back with its whole text in one field ready to post, as well as its setup and delivery separately, so a one-line joke and a two-part joke can be handled the same way.
• Get several jokesGet up to ten jokes in one request, with the same filters as a single draw. No joke is returned twice in one batch. The service changes the shape of its answer when you ask for more than one joke; this always returns a list, so a step that reads it does not break when someone raises the count from one to two. Fewer jokes than asked for means that is all that matched, which is reported rather than treated as a failure.
• Get a joke by idGet one specific joke by its id, for re-reading a joke an earlier step found or for pinning a workflow to a joke you have chosen. Ids are positions in a language's own joke file, so the same id is a different joke in each language and the language travels with it.
• Search jokesFind jokes containing a word or phrase, anywhere in the joke and whatever its case. Use it to pull a joke that fits an occasion, or to check whether the catalogue has anything about a subject at all. Nothing matching is an empty result here rather than a failed step, so a workflow can decide what to do about it. The service returns up to ten matches drawn at random from however many there were, and never reports how many that was, so treat the result as a sample rather than the first page of a ranked list.
• Get the joke of the dayPick one joke for a day and keep picking that same joke all day. The service draws at random on every call, so a daily post that retries, or that two people run an hour apart, tells two different jokes for the one day. This one is fixed by the date: the same date always gives the same joke, and tomorrow gives a different one. It honours the same filters as a normal draw, so a safe joke of the day stays safe. Add a seed of your own to have two workspaces tell different jokes on the same day.
• Get joke categoriesList the joke categories the service currently serves, and the aliases that resolve onto them. Worth having because a joke's own record says its category is Misc while a person filtering will as happily type Miscellaneous, and every category field in this integration takes either.
• Get blacklist flagsList the flags a joke can carry, which are the flags a filter can exclude. Run it before building a filter so a workflow excludes everything it means to, rather than the four someone remembered.
• Get languagesList the languages jokes exist in, with how many jokes each holds and how many of those are safe, alongside the languages the service translates its error messages into. These are two different sets, which is the point of this action: Italian and Russian are real JokeAPI languages with no jokes in them at all, so a workflow that picks a language off the wrong list gets nothing and cannot tell why.
• Get a language codeTurn a language name into the two-letter code the other actions take. The service searches its language list loosely, so German, Deutsch and a near miss all resolve. The answer also says whether that language has any jokes, because a code on its own does not: it resolves Italian quite happily, and there are no Italian jokes.
• Get API infoReport what the service currently holds: its version, how many jokes there are, which categories, flags, types and formats exist, which joke ids are valid in each language and how many of those jokes are safe. This is the action to run before filtering, and the one that says whether an id range you want to use exists at all.
• Check service statusCheck whether JokeAPI is answering, how quickly, and how much of this minute's request budget is left. This is the one action that reports a problem as a value instead of failing the step, so a workflow can decide what to do about it. Being throttled, being down for maintenance and being refused by the service's CDN are told apart, because they call for different responses: the first clears within a minute, the second is tracked on the service's status page, and the third usually clears on its own.

No actions match that.