Installing an API client 

Direct script tag 

Gadget generates JavaScript code directly for your web frontend for straightforward integrations into websites. This <script/> tag method is the simplest and easiest way to get started using openai-screenwriter-tutorial-v2's API.

<script src="https://openai-screenwriter-tutorial-v2--development.gadget.app/api/client/web.min.js"></script>
<script>
  // create a new instance of the Gadget client at the `api` global connected to the Development environment
  window.api = new Gadget()
</script>

Once the above script tags have been included, you can access api to start making API calls! The generated code is cached on Gadget's CDN and is safe to include directly on high-traffic web properties like a Shopify store or a WordPress blog. We recommend this method of installation for websites when possible, as the client will update without you having to make changes to your code or manually sync with Gadget.

const results = await api.query(`
  query GetSomeData {
    # start writing GraphQL queries!
  }
`);

const results = await api.query(`
  query GetSomeData {
    # start writing GraphQL queries!
  }
`);

<script src="https://openai-screenwriter-tutorial-v2.gadget.app/api/client/web.min.js"></script>

React 

To use your Gadget client with React, you must install both React and the React bindings.

npm install @gadgetinc/react @gadget-client/openai-screenwriter-tutorial-v2

yarn add @gadgetinc/react @gadget-client/openai-screenwriter-tutorial-v2

Creating the client for React apps 

Once your @gadget-client/openai-screenwriter-tutorial-v2 package has been installed, create an instance of the Client class in a shared file somewhere:

This client will be shared by all your React components, so it's best to put this code outside any particular component, or within one of your root components where it can be shared. Be sure to select the appropriate authentication mode for your use case.

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client({
  authenticationMode: { browserSession: true },
});

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client({
  authenticationMode: { browserSession: true },
});

export const api = new Client();

Providing the React client 

Once your client is installed and set up, you must provide an instance of your client to your React components using the Gadget React provider. Instantiate the client using the instructions above, and then pass it to the api prop of the provider.

1import { Provider } from "@gadgetinc/react";
2import { api } from "shared/api";
3const App = () => (
4  <Provider api={api}>
5    <YourRoutes />
6  </Provider>
7);

1import { Provider } from "@gadgetinc/react";
2import { api } from "shared/api";
3const App = () => (
4  <Provider api={api}>
5    <YourRoutes />
6  </Provider>
7);

Node module 

To work with the Gadget API server-side from node.js, or to include generated Gadget code in a larger product with a JavaScript bundler like Webpack or Parcel, you can install the npm module for openai-screenwriter-tutorial-v2. Installing the openai-screenwriter-tutorial-v2 API client works just like any other NPM package and works with any NPM-compatible package manager.

This installation method works by installing the package from Gadget's own npm registry, where we host, version, and update your API client.

To install the openai-screenwriter-tutorial-v2 node package, register the Gadget NPM registry for the @gadget-client package scope:

npm config set @gadget-client:registry https://registry.gadget.dev/npm

Then, install the openai-screenwriter-tutorial-v2 client package using your package manager:

npm install @gadget-client/openai-screenwriter-tutorial-v2

yarn add @gadget-client/openai-screenwriter-tutorial-v2

As you make changes to openai-screenwriter-tutorial-v2, Gadget will publish new versions of your API client to its NPM registry. See Client Regeneration for more details on when updates are necessary.

To indicate to hosting platforms where to find the @gadget-client/openai-screenwriter-tutorial-v2 package, create an .npmrc file in the web/frontend directory with the following content:

registry=https://registry.npmjs.org/
@gadget-client:registry=https://registry.gadget.dev/npm

After installation, the client module can be required and a client instance created:

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client();

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client();

The api object is now ready for use!

const results = api.query(`
  query GetSomeData {
    # start writing GraphQL queries!
  }
`);

const results = api.query(`
  query GetSomeData {
    # start writing GraphQL queries!
  }
`);

Gadget also makes a tarball of the latest version of the API client available for download.

https://openai-screenwriter-tutorial-v2--development.gadget.app/api/client/tarball

Connecting to a specific environment 

Gadget backend apps support two different versions of the application: a Development version for developers constructing and testing their application, and a Production version for end users. The API client can be configured to connect to either environment by passing the environment option to the client constructor.

For example, we can always connect to the Development environment by passing the string "Development" to the environment option:

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client({
  environment: "Development",
});
// all API calls will be made to the Development environment

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client({
  environment: "Development",
});
// all API calls will be made to the Development environment

You can also explicitly connect to the Production environment by passing the string "Production" to the environment option:

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client({
  environment: "Production",
});
// all API calls will be made to the Production environment

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client({
  environment: "Production",
});
// all API calls will be made to the Production environment

or you can select which environment to connect to based on an environment variable

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client({
  environment: process.env["IS_PRODUCTION"] == "true" ? "Production" : "Development",
});

import { Client } from "@gadget-client/openai-screenwriter-tutorial-v2";
export const api = new Client({
  environment: process.env["IS_PRODUCTION"] == "true" ? "Production" : "Development",
});

By default, if no environment is specified the client will set the environment based on NODE_ENV if it is set.

Type safety 

The default TypeScript types are included in the openai-screenwriter-tutorial-v2 client package, requiring no additional installation steps to utilize them.

The openai-screenwriter-tutorial-v2 client ensures comprehensive type safety for high-level JavaScript functions like find and findMany, providing reliable types for API requests and responses. When using the select option to retrieve specific fields, the client maps the selected fields to the GraphQL schema, returning an object with the chosen fields and their correct types, as demonstrated in this TypeScript example fetching the id and name of a User model.

1const users = await api.user.findMany({ select: { id: true, name: true } });
2
3console.log(users[0].id);
4// will typecheck just fine as the type `string`, and log an ID string for the user
5console.log(users[0].name);
6// will typecheck just fine as the type of the `name` field as set in the Gadget Editor, and log the value
7console.log(users[0].email);
8// will fail the typecheck, as the `email` key was not included in the `select` option

1const users = await api.user.findMany({ select: { id: true, name: true } });
2
3console.log(users[0].id);
4// will typecheck just fine as the type `string`, and log an ID string for the user
5console.log(users[0].name);
6// will typecheck just fine as the type of the `name` field as set in the Gadget Editor, and log the value
7console.log(users[0].email);
8// will fail the typecheck, as the `email` key was not included in the `select` option

Client regeneration 

The API client source code changes as you make changes to your Gadget backend.

Because the web clients are generated on-demand, we recommend using these clients where possible. By sourcing it directly as a <script/> tag, your browser will fetch the most up-to-date code each time a page loads, and you'll always be working with a client guaranteed to work with the current API for the application. To preserve this fast refresh behavior, it is best to not cache the client with other infrastructure (like a proxy) or to save it to disk.

Updating the npm package 

Gadget's npm registry will publish new versions of your package as changes are made to your Gadget app. You must install the new version of the API client package locally to reflect changes made in the Gadget editor. If a new model is added or removed, it will be available in the generated clients right away, but you need to update your local dependency version to fetch these changes.

To upgrade using npm, run:

npm install @gadget-client/openai-screenwriter-tutorial-v2@latest

To upgrade using yarn, run:

yarn add @gadget-client/openai-screenwriter-tutorial-v2@latest

Updating to only the latest production version 

Upgrading to the @latest version of the npm package will install the latest version for the development environment. To upgrade to the latest version of the production environment only, you can install the @latest-production version. This version of the API client will only reflect changes that have been deployed to production, and will not show changes that are in development but not yet deployed.

To upgrade to only the latest deployed changes using npm, run:

npm install @gadget-client/openai-screenwriter-tutorial-v2@latest-production

To upgrade to only the latest deployed changes using npm, run:

yarn add @gadget-client/openai-screenwriter-tutorial-v2@latest-production

Once changes are deployed, the @latest and @latest-production versions of the packages will match.

Best-effort functionality 

If your client version is out of date, that version still makes its best effort to continue to work with new versions of the data model and GraphQL schema. If a breaking change is made the client may try to make requests that no longer work. This can always be remedied by updating any cached source code using npm or yarn, or by copying new files sourced from Gadget's web.js endpoint.

The following changes will break an out-of-date openai-screenwriter-tutorial-v2 client and cause it to issue invalid requests:

• changing a model's apiIdentifier
• changing an action's apiIdentifier
• removing an action
• removing a model completely