How to Setup Cloudflare Workers for Serverless Compute

SlinkyIO — Thu, 10 Sep 2026 20:43:00 +0000

I’ve spent a long time looking for a decent yet simple cloud compute solution, and I think Cloudflare Workers solved my issue. I’ve been working on an API project that needs some persistent compute infrastructure and database access. Cloudflare’s generous free tier and hosted worker solution was the starting point I needed. So how did I set the project up?

The Stack

HonoJS for my web framework. Hono is lightweight, fast, and contains the routing features I need for the API.
Supabase for the backend database. One table holds the data that the API fetches, and the supabase-js library makes building the routes straightforward.
Cloudflare Workers. Serverless and persistent hosting means the API is operational 24/7 with limited infrastructure management on my end. Logging and observability are baked in, and a GitHub integration means I can deploy updates with one commit.

Setup

First, create a new project using npm create cloudflare@latest - my-first-worker
Follow the prompts to set up the project.

From here, development work is normal Typescript/Javascript. Define routes, write code, etc. Here is a simple generic example of what a worker route serving an API could look like:

const app = new Hono<{ Bindings: Bindings }>()

function makeSupabaseClient(c: AppContext) {
  return createClient(
    <<yoururlhere>>
    <<yourkeyhere>>
  )
}

async function respond<T>(
  c: AppContext,
  query: PromiseLike<{ data: T | null; error: { message: string } | null }>
) {
  const { data, error } = await query
  c.header('cache-control', 'public, max-age=300')
  if (error) {
    return c.json({ error: error.message }, 500)
  }
  return c.json(data ?? [])
}

app.get('/example/:param', async (c) => {
  const supabase = makeSupabaseClient(c)
  const { param } = c.req.param()
  return respond(c, paginate(c, supabase.from('somedatabase').select(mycolumns).eq('param', param)))
})

I’ve created some reusable functions here because my service has multiple routes, but a simpler implementation likely doesn’t need that.

Wrangler and Deployments

Wrangler is Cloudflare’s CLI tool for developing code for workers and for performing manual deployments. If Wrangler wasn’t installed during project creation, install it with this command: npm i -D wrangler@latest

If you’re planning on committing your work to GitHub, there is a Cloudflare worker integration that automates deployments based on pushes to the main branch. These docs describe how to connect a worker deployment to a GitHub repository. The setup is very straightforward, and I’d highly recommend it if you’re going to keep your code on GitHub anyway.

Aside from the automated deployments, you can use this command to manually deploy a change if you need to: npx wrangler deploy

What can you use these workers for?

That’s a great question, and the answer is a lot. The basic workers can be used for anything a standard web worker would do. Fetching API content, data processing, and concurrency are all common use cases. Cloudflare also has a more robust implementation called Pages for serverless full-stack applications and a Workers AI for creating and hosting custom AI implementations.

Key Features

The worker’s dashboard has a ton of great additions. Built-in logging and observability means you don’t have to implement Jaeger or Prometheus for observability. Other features include support for custom domains, automated email sending, and Zero Trust integrations.

Additionally, if you like to use AI, Cloudflare provides a connector for testing and monitoring workers. I’ve tried using the Claude connector (I’m sure the other LLMs have it too), and it was able to provide me with information about the status of the worker. I’d recommend still having a hand in the development and deployment of the worker (not because the connector isn’t capable, but because you never know what code AI tools will write), but this could still be useful as some kind of automated status update or a more robust monitoring tool outside of the platform.

Conclusion

If you’re like me and have been looking for an accessible and free (initially) solution to cloud compute for your projects, I highly recommend looking into these Cloudflare Workers. The generous free tier creates a low barrier to entry for anyone wanting to try the service out over a weekend. Please let me know if you try these out and what you use them for.