TechPrufer
Get started
Vercel
Vercel logoVercel · Feature trail

Vercel Services

8 posts · May 2023 – Jun 2026

What this trail covers

Vercel Services allows running multiple frameworks (frontend, backend, agents, etc.) within a single Vercel Project, enabling atomic deployments, shared preview deployments, and internal service communication. Services are configured via `vercel.json` and can communicate internally via service bindings, with traffic staying on Vercel's network. The platform supports zero-configuration deployment for various Python and TypeScript frameworks, leveraging Fluid Compute with Active CPU pricing. It in Vercel's platform now extends zero-configuration deployment to backends, supporting popular Python and TypeScript frameworks like FastAPI, Flask, Express, Hono, NestJS, and Nitro. This Framework-Defined Infrastructure (FDI) approach automatically provisions and scales infrastructure based on framework conventions, eliminating the need for manual configuration files or Docker images. Backends are billed using Active CPU Pricing on Fluid Compute, meaning users only pay for active execution time, not idle waiting. For long-running processes, the Workflow Development Kit (WDK) provides durable orchestration and observability. New backend AI templates are also available, demonstrating patterns like RAG pipelines.

All Vercel trails
  1. Vercel Services: Run full stack on VercelIntroduces Vercel Services, a feature allowing multiple frameworks (frontend, backend, etc.) within a single Vercel Project. This enables atomic deployments, shared preview deployments, and internal service communication via service bindings. Configuration is done via `vercel.json`. Services can talk to each other internally without leaving Vercel's network. The post highlights zero-configuration deployment for various frameworks, integration with Fluid Compute and Active CPU pricing, Vercel Sandbox for isolated agent compute, WebSocket support, Vercel Connect for secure external access, Vercel Marketplace for databases, Vercel Queues/Workflows/Cron for background jobs, and Secure Compute/static IPs/VPC peering for private connections. Functions now support up to 30 minutes duration on Pro/Enterprise and 500MB dependency size for Python.
  2. Zero-config backends on Vercel AI CloudIntroduces zero-configuration deployment for Python (FastAPI, Flask) and TypeScript (Express, Hono, NestJS, Nitro) backends on Vercel AI Cloud. Leverages Framework-Defined Infrastructure (FDI) to automatically provision and scale infrastructure based on framework conventions. Extends Fluid Compute with Active CPU Pricing for intermittent workloads, billing only for active execution time. Introduces the Workflow Development Kit (WDK) for durable orchestration of long-running backends and background agents. Offers new backend AI templates for patterns like RAG pipelines.
  3. Update regarding Vercel service disruption on October 20, 2025This post details Vercel's response to the AWS us-east-1 outage, highlighting the resilience of its multi-region architecture for serving production traffic. It explains how Vercel's stack is designed to survive the loss of any single region and details the cascading failure that affected its global caching infrastructure, leading to temporary disruption of static file serving. The post also describes a secondary incident impacting Vercel's control plane due to a feature flag provider outage, which affected the dashboard, APIs, and builds but not production traffic serving. Mitigations and restoration efforts for both incidents are outlined, emphasizing the importance of regional fault tolerance and the need for further improvements to prevent any request drops.
  4. Server rendering benchmarks: Fluid Compute and Cloudflare WorkersThis post introduces benchmarks comparing Vercel's Fluid compute against Cloudflare Workers for server-side rendering. It highlights Fluid compute's advantages in performance (1.2-5x faster) and consistency due to in-region deployment, full Node.js/Python runtime compatibility, and configurable concurrency/resource allocation. The benchmarks demonstrate significant improvements for frameworks like Next.js, React SSR, and SvelteKit, as well as compute-bound tasks.
  5. Vercel + Sanity: Innovating on a fast, collaborative Web with features like Visual EditingIntroduces Visual Editing as a joint solution with Sanity, enabling a one-click path from Vercel's Preview Deployment interface to Sanity Studio for content editing. This enhances collaboration between developers and marketers by allowing visual editing of content directly on the frontend, regardless of where it's reused. It also highlights the composable web architecture enabled by Vercel and Sanity, citing examples like Loom, and emphasizes the benefits of decoupling frontend and backend for developer workflow, user experience, and content management.