Vercel Launches Python in General Availability: Six SDKs and FastAPI Support

Vercel launched Python in general availability. Developers now have the ability to deploy Python applications on the same platform as their JavaScript frontend, with automatic framework detection, Fluid compute for scaling, and six specialized Python SDKs for integration with cloud infrastructure.

Principles and Architecture

Vercel followed three key principles when developing Python support. First, APIs use type annotations as is standard in the Python ecosystem, and work with CPython on Linux for compatibility with packages and tools. Second, a unified workflow for development, preview deployments, and production of frontend (Next.js, Vite) and backend. Third, automatic framework detection and configuration.

The platform analyzes dependency files (pip, poetry, etc.) and automatically detects frameworks: FastAPI, Flask, Django, FastHTML — configuring the build with request routing. Services allows deploying frontend and backend in a single deployment, running `npm start` brings both parts up locally, preview deployments show both services together, production is rolled out synchronously.

Six Python SDKs for Cloud Infrastructure

Vercel provides specialized SDKs for working with cloud services:

• AI SDK for Python — stream and generate model responses, call tools, build agents • Sandbox — create and manage isolated environments for code execution • Workflows — define long-lived workflows with support for retry attempts and execution state persistence • Queues — send and receive messages for background jobs and notifications • Blob — upload and retrieve files from object storage • External Credentials — obtain credentials for calls to external APIs

Computing runs on Fluid compute — infrastructure that manages parallel requests and automatically scales with traffic. Static files go to CDN via Services.

Demo: Collaborative Application for Python Notebooks

The Vercel team built a platform for creating and viewing Python notebooks with an AI assistant that demonstrates the integration of all components. Architecture: Vite frontend, FastAPI backend, PostgreSQL for storing notebook sources and chat history, Services for unified deployment with shared CDN.

When a notebook is opened, FastAPI creates or reuses a Vercel Sandbox with Jupyter for each user (isolated environment). The browser connects directly to Jupyter, Mounted Drives provide Python dependencies and workspace files. The browser periodically sends state to FastAPI. The backend saves the source in PostgreSQL, renders changes to HTML, and uploads the result to Blob. Published notebooks are read through CDN without running the Python kernel.

FastAPI sends a message to Vercel Queues upon publication. All backend instances with connected browsers receive a notification and broadcast updated lists via WebSockets — users see changes in real time without reloading.

The AI agent uses AI SDK for Python and executes within Vercel Workflows, which preserve state and retry failed steps. Chat history is in PostgreSQL, code editing and execution tools are connected through the browser, results are returned via FastAPI to resume the workflow.

Investments in the Python Community

In December 2025, the Gel Data team led by Elvis Pranskevichus joined Vercel to develop Python support and open-source work. Vercel sponsors the Python Software Foundation and finances annual CPython maintenance by developer Serhiy Storchaka, who has written over 600 commits to the CPython main branch in 2026.

The company is a sponsor of PyCon US 2026 and EuroPython 2026, where the team met with developers and heard about practical problems. These conversations determine the platform's development priorities.