The fal.ai
JavaScript Client Library provides a seamless way to interact with fal
serverless functions from your JavaScript or TypeScript applications. With built-in support for various platforms, it ensures consistent behavior across web, Node.js, and React Native environments.
[!WARNING] This dependency was deprecated in favor of the official 1.0.0 release, renamed to
@fal-ai/client
. Please update your dependencies to the new package.
Before diving into the client-specific features, ensure you've set up your credentials:
import * as fal from "@fal-ai/serverless-client";
fal.config({
// Can also be auto-configured using environment variables:
// Either a single FAL_KEY or a combination of FAL_KEY_ID and FAL_KEY_SECRET
credentials: "FAL_KEY_ID:FAL_KEY_SECRET",
});
Note: Ensure you've reviewed the fal.ai getting started guide to acquire your credentials and register your functions. Also, make sure your credentials are always protected. See the ../proxy package for a secure way to use the client in client-side applications.
The fal.run
method is the simplest way to execute a function. It returns a promise that resolves to the function's result:
const result = await fal.run("my-function-id", {
input: { foo: "bar" },
});
The fal.subscribe
method offers a powerful way to rely on the queue system to execute long-running functions. It returns the result once it's done like any other async function, so your don't have to deal with queue status updates yourself. However, it does support queue events, in case you want to listen and react to them:
const result = await fal.subscribe("my-function-id", {
input: { foo: "bar" },
onQueueUpdate(update) {
if (update.status === "IN_QUEUE") {
console.log(`Your position in the queue is ${update.position}`);
}
},
});
The client library offers a plethora of features designed to simplify your serverless journey with fal.ai
. Dive into the official documentation for a comprehensive guide.