Cloudflare's Python Workers reach general availability with native bindings and framework support

Python Workers reach general availability on Cloudflare

Cloudflare has moved Python Workers to general availability, a milestone the company says makes Python a first-class and fully supported language across its developer platform. According to the Cloudflare blog, the feature arrived two years after its introduction, and GA status means Python code, libraries and familiar design patterns can now be wired directly into services such as Workers AI, R2, D1, Hyperdrive, Durable Objects, Queues and Workflows. A Python Worker can also be created from inside another Worker via Dynamic Workers.

A Wasm-compiled interpreter, two years in

The runtime is not new plumbing. Cloudflare explains that Workers has supported WebAssembly since 2018, which made it a natural environment for a Wasm-compiled Python interpreter — the team built on Pyodide to cover a wide range of Python applications quickly. The multi-year effort was aimed at making Python on Workers feel as simple and performant as Python anywhere else, with the ambition of scaling Python apps without capacity limits. Many developers were already building on the preview; the capabilities highlighted now are production-ready for everyone.

Platform bindings without JavaScript glue

A major friction point in the preview era was interoperability. Passing a Python dictionary into a Cloudflare Queue, for example, required explicitly converting it into a JavaScript object at the RPC boundary, using Pyodide's to_js helper plus a dictionary converter. Cloudflare says this forced Python developers to keep the JavaScript environment in mind while writing Python, and was a recurring source of errors — for humans and AI coding agents alike.

That conversion work has been absorbed into the Workers runtime and the Python SDK. Sending a plain Python dictionary to a Queue binding now works as-is, with no JavaScript written by the developer, which Cloudflare presents as the practical meaning of first-class support.

FastAPI, Django and Flask support

Python Workers now ship built-in connectors for web frameworks. Async applications such as FastAPI can be mounted through the workers.asgi package, while synchronous apps such as Django use workers.wsgi, typically via a one-line entrypoint declaration. In a conventional deployment these apps sit behind servers like uvicorn or gunicorn, but on Workers the platform itself acts as the web server, relying on Cloudflare's global network for load balancing and scaling.

Under the hood, the connectors form a lightweight bridge that translates incoming native JavaScript requests into the standard WSGI or ASGI structures Python applications expect, then pipes responses back out with little added overhead. Because the bridge targets those standard interfaces, it works with any WSGI- or ASGI-compatible framework, not only the three highlighted ones.

TCP sockets and Hyperdrive for databases

The release also closes a significant gap: TCP networking. Python database drivers such as asyncpg and aiomysql depend on the standard library's socket module, which ordinarily issues POSIX system calls to the host operating system. Inside a WebAssembly sandbox those networking syscalls are placeholder implementations that always fail, so opening a standard connection was previously impossible.

Cloudflare's fix was to implement the socket system calls on top of the Workers connect API. When a driver tries to open a TCP connection, the custom layer translates ordinary Python socket operations — connecting, reading bytes — into the JavaScript calls the runtime understands. Since the translation happens at the syscall level, the drivers themselves need no modifications. That bridge is what enables Hyperdrive support for PostgreSQL and MySQL: the binding is configured in Wrangler, and drivers connect using the host, port and credentials Hyperdrive exposes.

Why it matters

Serverless and edge platforms have largely been JavaScript- and TypeScript-first, pushing the large population of Python developers — particularly in AI and data work — toward containers or conventional infrastructure. GA on Workers means a familiar Python stack, from FastAPI routes to standard database drivers, can now run on a globally distributed runtime with no servers to manage. It is also a meaningful validation of WebAssembly as a server-side technology: a mainstream platform shipping a production-ready, non-JavaScript language runtime on Wasm, complete with syscall-level networking workarounds. For teams already on Cloudflare, removing cross-language glue code eliminates a whole class of bugs; for the wider market, it raises the pressure on other serverless providers to treat Python as a first-class citizen rather than an afterthought.