Skip to content

How It Works


The internal /api/__kirby__ server route proxies KQL requests to the Kirby instance.


The proxy layer will not only pass through your API's response body to the response on the client, but also HTTP status code, HTTP status message and headers. This way, you can handle errors just like you would with a direct API call.

Detailed Answer

The useKql and $kql composables will initiate a POST request to the Nuxt server route /api/__kirby__ defined by this module. The KQL query will be encoded in the request body.

The internal server route then fetches the actual data for a given query from the Kirby instance and passes the response back to the client. Thus, no KQL requests to the Kirby instance are initiated on the client-side. This proxy behavior has the benefit of omitting CORS issues, since data is sent from server to server.

During server-side rendering, calls to the Nuxt server route will directly call the relevant function (emulating the request), saving an additional API call. Thus, only the fetch request from Nuxt to Kirby will be performed.


Query responses are cached (with a key calculated by hashing the query) and hydrated to the client. Subsequent calls will return cached responses, saving duplicated requests. Head over to the caching section to learn more.

Released under the MIT License.