Cloudflare · Workers Platform

What Are Cloudflare Workers? Edge Serverless Explained

Last verified: June 17, 2026  ·  Format: Breakdown

Cloudflare Workers is a serverless platform that runs your code on Cloudflare's global edge network, close to your users, without you managing any servers. If you have been wondering what they are in practical terms, picture writing a small function, deploying it once, and having it run from data centers in cities around the world automatically. What makes the platform distinctive is the execution model: instead of spinning up a container or a virtual machine for each request, Workers run on lightweight V8 isolates, the same engine technology that powers your browser's JavaScript. The result is fast startup, no cold starts, and a billing model where you pay for CPU time rather than idle time.

This breakdown is plain and practical. We start with what the platform actually is, then explain the isolates model that sets them apart from containers, walk through free and paid pricing with exact figures, cover the languages and the wider platform products you can build with, and finish with when to reach for Workers and how they fit alongside the big clouds. Product and pricing details below are drawn from Cloudflare's own documentation and were checked on June 17, 2026; treat them as a starting point and confirm current figures with Cloudflare.

What Are Cloudflare Workers?

The clearest way to answer what Workers are is this: they are a serverless way to run code on Cloudflare's network rather than on a server you provision and maintain. You write a function, deploy it to the platform, and Cloudflare runs it on demand at locations near your users. There are no machines to size, no clusters to scale, and no operating systems to patch; the network handles execution for you.

Cloudflare runs Workers on its edge, the data centers it operates in more than 330 cities worldwide (a figure Cloudflare reports). Because your code runs close to wherever a request originates, requests are handled with low latency instead of traveling to one distant region. And because Workers start almost instantly, there is no cold start penalty, the lag you can feel on some serverless platforms when an idle function has to wake up before it can respond.

One way to place Workers in context: Cloudflare is a connectivity cloud built around its network, and Workers are the part of that platform where you build new applications on the network itself. If you want the wider picture of the company and its other products first, read what Cloudflare is, then come back here for the developer platform. For the full landscape across vendors, start at the Cloud Tools hub.

How the Isolates Model Works

The feature that most defines Workers is what they run on. Rather than packaging each function in its own container or virtual machine, Workers execute inside V8 isolates, the lightweight sandboxes that the V8 JavaScript engine uses to keep code separate. Many isolates can share a single process, which is what lets the platform start your code in a fraction of the time a container would take to boot.

Isolates versus containers

The practical difference comes down to overhead. A container or virtual machine bundles an operating system and runtime, so it has to be provisioned and started before your code runs; when one has been idle, that start-up shows up as a cold start. An isolate, by contrast, is far lighter and is kept ready, so a Worker can begin handling a request almost immediately. This is why Cloudflare can credibly say Workers have no cold starts, and why they suit short, latency-sensitive tasks that run very frequently.

The trade-off built into this model is the unit you are billed on. Because isolates are so light, Cloudflare charges for the CPU time your code actually uses, not for the wall-clock time a request is open or for idle time while a function waits. If your Worker spends most of a request waiting on a network call, that waiting does not run up CPU charges. That is a different shape of billing from platforms that charge by memory multiplied by total duration, and it rewards code that does short bursts of work.

Workers Pricing: Free and Paid

Workers pricing comes in two tiers, and the free one is genuinely useful for learning and small projects. The figures below are Cloudflare's published numbers, checked on June 17, 2026; because pricing and limits change, confirm the current details on Cloudflare's own Workers pricing page before you commit.

The headline to take away is that you pay for what your code computes, not for keeping it available. On the Free plan you get 100,000 requests per day and 10 ms of CPU per invocation at no charge for duration. Workers Paid starts at a $5 per month minimum that includes 10 million requests and 30 million CPU-milliseconds, with overage at $0.30 per additional million requests and $0.02 per additional million CPU-ms. There is no separate egress or bandwidth charge in either tier.

Languages and Platform Products

Workers are not limited to JavaScript, and they are rarely used alone. The platform supports several languages and surrounds the compute layer with storage and data products you can bind directly to your code, which is what lets you build complete applications on the network.

Supported languages

You can write Workers in JavaScript and TypeScript, which are the most common choices, as well as Python, Rust, and C and C++. The platform also runs WebAssembly (Wasm), so code compiled from other languages to Wasm can execute on the edge too. This range means you are not forced into one ecosystem to use the platform.

The wider Workers platform

Around the compute layer, Cloudflare offers a set of products that you connect to your Workers to add state, storage, and data:

• Workers KV: a global, low-latency key-value store for configuration and cached data.
• R2: object storage that Cloudflare markets with zero egress fees, so you are not charged to read your data back out.
• Durable Objects: stateful coordination primitives for things that need a single, consistent point of state.
• D1: a serverless SQL database for relational data.
• Queues: message queuing for decoupling and background processing.
• Hyperdrive: connection pooling and acceleration for your existing databases.
• Workers AI: running machine-learning models on Cloudflare's network.
• Vectorize: a vector database for embeddings and AI-powered search.

When to Use Cloudflare Workers

We have covered what Workers are and how they are built; the next question is when to reach for them. Because the model rewards short, frequent, latency-sensitive work and the Free plan lowers the barrier so far, the platform suits a broad set of jobs, but some fit better than others.

How Workers Fit Alongside AWS, Azure, and GCP

A common question about Workers is whether they replace serverless functions on the big clouds, such as AWS Lambda. For most workloads the honest answer is that Workers complement the hyperscalers rather than replacing them. Workers are purpose-built for edge serverless: short, latency-sensitive code running close to users. The heavy application backends, large managed databases, and general-purpose compute that many teams run still tend to live on a hyperscaler or your own hosting.

The reason is the difference in shape. AWS, Microsoft Azure, and Google Cloud are general-compute clouds where you provision a deep catalog of infrastructure and run your application there. Workers focus on the network edge and lightweight compute, with the isolates model optimized for fast startup rather than long-running, resource-heavy jobs. The platform lets you build real applications, but it is not a drop-in substitute for a hyperscaler's full compute and data stack.

In practice, many teams use both: a hyperscaler for the heavy backend and Workers for edge logic, fast APIs, and request-time work in front of it. To see how an equivalent hyperscaler function service is structured, read what AWS Lambda is, and to understand where responsibilities sit when you run on a managed platform, see the cloud shared responsibility model.

Honest Trade-offs

No answer to what Workers are would be complete without the trade-offs. Workers are a strong default for edge serverless, and the points below are not reasons to avoid them. They are reasons to adopt them with clear eyes.

Frequently Asked Questions