Custom API (beta)

Deploy a custom API on Nhost. Also called: Lambda, Serverless Functions, Cloud Functions etc.

To deploy the custom API you must:

  1. Init your project with the nhost-cli: nhost init
  2. Connect your repository to your Nhost project under Settings -> Git.
  3. Have an api/ directory at the root of your project.

Your API will be automatically deployed for every push to the master or the main branch of your repository.

#Example code

module.exports = (req, res) => {
  const { name = 'World' } = req.query;
  res.status(200).send(`Hello ${name}!`);


The folder structure inside the api/ folder will resolve the endpoints.

If you're deploying your frontend to Vercel you can add api to .vercelignore so the API-folder does not conflict with Vercel's deployment.

Example paths and endpoints

Given this folder and file structure:

| |____index.js
| |____users
| | |____index.js
| | |____active.js
| |____my-company.js

The following endpoints will be available: - (api/index.js) - (api/users/index.js) - (api/users/active.js) - (api/my-company.js)

#Environment variables

System level environment variables are available:

NHOST_HASURA_ADMIN_SECRET - (Hasura admin secret)
NHOST_HBP_URL - (HBP URL for auth and storage)
NHOST_WEBHOOK_SECRET - (Random secret string. Can be used to authenticate requests from Hasura to the Custom API)

You can access each environment variable via process.env.<name> ex: process.env.NHOST_HASURA_URL.

All your production environment variables added under Settings -> Environment variables are also added to the custom API.

For local development all system environment variables will point to the local development URLs.


Logs are not available yet.

#Local development

Use the Nhost CLI's nhost dev to start the Custom API locally (together with a full Nhost backend).

#TypeScript support

TypeScript is supported.

import { Request, Response } from "express";

export default (req: Request, res: Response) => {
  const { name = 'World' } = req.query;
  res.status(200).send(`Hello ${name}!`);

#Technical details

We use Google Cloud Run to deploy the Custom API. For now we only allow JavaScript. In the future you will be able to bring your own Dockerfile with any language.