Standard Library

api

Make outbound HTTP requests. Backed by libcurl, with JSON parsed into native objects.

Import:

import "api";

curl

Send a request and get the parsed response. The first argument is a URL; the second is an options object.

object res = curl("https://dogapi.dog/api/v2/facts?limit=1", {});
res->data[0]->attributes->body | print;

Options

All options are optional:

methodGET, POST, PUT, DELETE, etc. Default GET.

headersAn object of header name to value.

bodyAn object that will be sent as JSON.

POST with a JSON body

curl("https://api.example.com/items", {
    method: "POST",
    headers: {
        Authorization: "Bearer " + token
    },
    body: {
        name: "widget",
        price: 9.99
    }
}) | print;

Calling an LLM

Most chat APIs follow the OpenAI shape, so a single curl call works against many providers:

import "dotenv";

string url = envvar("AI_URL");
string model = envvar("AI_MODEL");
string key = envvar("AI_API_KEY");

curl(url, {
    method: "POST",
    headers: { Authorization: "Bearer " + key },
    body: {
        model: model,
        messages: [
            { role: "system", content: "You are friendly." },
            { role: "user", content: "Hi!" }
        ]
    }
}) | print;