Summary
- Python Workers is generally available after two years of development, making Python a fully supported language on Cloudflare’s developer platform.
- Developers can run FastAPI, Django, and Flask while connecting directly to services including Workers AI, D1, R2, Hyperdrive, Queues, and Durable Objects.
- Expanded networking and WebAssembly package support brings more database and AI libraries into the Python runtime.
Cloudflare has made Python a fully supported language on its Workers platform, widening the serverless service beyond its JavaScript roots as developers increasingly use Python for web applications, data processing, and artificial-intelligence workloads.
Python Workers has moved into general availability two years after Cloudflare first introduced the capability, with support for mainstream frameworks including FastAPI, Django, and Flask. Applications can connect directly to other parts of the company’s developer platform, including Workers AI, D1, R2, Hyperdrive, Durable Objects, Queues, and Workflows.
The technical change is partly about removing boundaries between Python code and the JavaScript-based runtime underneath it. Developers previously had to convert Python objects explicitly when using some Cloudflare services, whereas the runtime and Python software development kit now handle that conversion automatically.
Cloudflare has also extended networking support so Python applications can use familiar database drivers and AI libraries that previously struggled inside its WebAssembly environment. The result is a more conventional Python development experience running on infrastructure distributed across Cloudflare’s network.
Serverless Python loses some special handling
Cloudflare runs Python through Pyodide, a version of the language compiled to WebAssembly, rather than placing a conventional Python server on every Workers instance. That design provides isolation and allows code to run inside the same serverless environment as other Workers applications, but it has historically created compatibility problems around networking and packages containing compiled extensions.
The general-availability release addresses several of those limitations. Cloudflare has built WSGI and ASGI connectors that allow existing Python web frameworks to interact with the Workers runtime, meaning a FastAPI, Django, or Flask application can retain a familiar application structure while Cloudflare handles incoming requests and scaling.
Database access has also become less unusual. Python libraries connecting to PostgreSQL and MySQL depend on low-level socket behaviour that was previously unavailable inside the WebAssembly sandbox. Support for that networking path now allows applications to use compatible database drivers through Cloudflare’s Hyperdrive service.
Those changes reduce the amount of platform-specific adaptation required when a development team already has Python expertise. Serverless services have long promised to remove infrastructure management while replacing part of it with their own runtime assumptions, supported languages, deployment patterns, and compatibility constraints. Broadening the ordinary Python tooling that works without modification reduces one element of that trade-off.
The package ecosystem remains a limitation. Python libraries containing native C, C++, or Rust extensions have to be compiled for WebAssembly before they can run in Python Workers, and Cloudflare acknowledges that not every package is available.
The company has been working upstream rather than maintaining an entirely proprietary package system. Its engineers proposed PEP 783, standardising a Python-on-WebAssembly platform called PyEmscripten, and have contributed to build tooling intended to help package maintainers publish compatible versions.
Python opens a clearer route into AI workloads
The language expansion also strengthens Cloudflare’s position in AI development because much of the machine-learning and agent ecosystem is built around Python. Libraries such as OpenAI’s client, LangChain, and Model Context Protocol tooling can now run in Python Workers, according to the company.
A development team can therefore combine third-party model APIs with Workers AI, use AI Gateway as an intermediary, store application state in Cloudflare services, and orchestrate background work without changing programming language merely to reach the surrounding infrastructure.
Cloudflare has published examples covering asynchronous AI pipelines, Model Context Protocol servers, retrieval-augmented generation, and persistent streaming connections. Those patterns put Python Workers into competition not only with conventional serverless platforms but also with managed environments built specifically around AI applications.
The architectural questions remain familiar. Developers still have to decide where data resides, which services become proprietary dependencies, whether workloads fit an edge-oriented execution model, and how applications behave when they rely on systems running outside Cloudflare’s own network.
Python support nevertheless removes one adoption barrier. A data or AI team can build against a language and libraries it already uses while infrastructure teams retain the operational characteristics of Workers, rather than rewriting the application around JavaScript or splitting execution between platforms.
Cloudflare says it still intends to improve memory efficiency, performance, and package coverage. General availability therefore does not make the runtime identical to a conventional Linux Python environment, but it moves the technology from an experimental compatibility layer towards a supported production option.
