Framework-Defined Infrastructure

6 posts · Mar 2023 – Sep 2026

What this trail covers

Vercel's platform is evolving towards self-driving infrastructure, leveraging AI agents to autonomously manage production operations, improve application code with real-world insights, and learn from production behavior. This builds upon the foundation of framework-defined infrastructure, where application code dictates infrastructure provisioning. The Vercel AI Cloud enables new primitives for AI workloads, and the Vercel Agent acts as a first responder for production operations and security, investigating and resolving issues.

All Vercel trails

Sep 2026State of agent skillsThis post analyzes the growth and adoption of agent skills, highlighting the rapid expansion of the skills.sh registry and the distribution of skill categories by supply and demand. It contrasts the technical focus of skill creation with the broader application of agents across various business functions. The post identifies agent workflows and automation as a highly installed category, indicating a strong demand for skills that improve how agents operate.
Jul 2026The Vercel CLI is Now Built for AI AgentsThe Vercel CLI is being re-architected to serve as a building primitive for AI agents. This involves exposing programmable interfaces for connecting to resources, creating sandboxes, inspecting logs, and debugging deployments. New CLI commands and features like `vc flags` for feature flag management and `vc curl` for testing deployments are introduced. The CLI will also support experimental native binaries and plugins for AI code assistants like Claude Code, Codex, and Grok, enabling agents to interact with Vercel's platform programmatically.
Jul 2026From init to deploy in 30 minutes - Ship 26 NYC WorkshopThis talk demonstrates the practical application of Stripe Projects, a Stripe CLI plugin that enables users and coding agents to provision hosting, databases, domains, and authentication. The workshop showcases a 30-minute hackathon where an agent, guided by prompts, initializes and configures these services, highlighting the potential for AI-assisted infrastructure provisioning and rapid application deployment. The focus is on the agent's ability to interpret prompts, define service requirements, and interact with the Stripe Projects CLI to set up a functional application stack.
Jul 2026Building full stack AI application with Vercel and AWS - Ship 26 NYC WorkshopThis workshop demonstrates building a full-stack AI application by integrating Vercel's platform with AWS services. It covers deploying AI models and applications, leveraging Vercel's infrastructure for frontend and backend components, and integrating with AWS for scalable AI workloads. The session highlights how Vercel's framework-defined infrastructure and AI Cloud primitives can be combined with AWS's robust services to create end-to-end AI solutions.
Jul 2026Opening Keynote - Ship 26 NYCThis keynote introduces the agentic infrastructure built for a platform where humans and agents build, deploy, and run websites, apps, and agents together. It highlights that over half of all deployments on Vercel now come from coding agents, and increasingly, those agents are shipping agents of their own. Key announcements include Fluid Compute 2.0, Docker, and Vercel Services, along with updates to the AI SDK, AI Gateway, Workflow SDK, Sandbox, Connect, and Chat SDK. The introduction of 'eve' and the Vercel Agent are also featured, alongside a discussion of over 100 agents in production and Vercel for Enterprise Apps and Agents.
Advertisement
Nov 2025Vercel: The anti-vendor-lock-in cloudThis post elaborates on the concept of Framework-Defined Infrastructure (FDI) as a core principle of Vercel's platform. It contrasts FDI with vendor-specific primitives found on platforms like AWS and Cloudflare, emphasizing how Vercel interprets framework code (e.g., Next.js, Remix, SvelteKit) to automatically provision necessary infrastructure without requiring Vercel-specific modules or APIs in the application code. The post highlights the portability benefits of FDI, enabling applications to run on Vercel or other platforms supporting the same framework conventions. It also discusses how FDI simplifies local development by using standard framework development servers instead of vendor-specific simulators. The post mentions the role of Build Adapters in formalizing the framework-platform contract and reiterates Vercel's commitment to open standards and portable code, citing examples of Next.js applications running outside of Vercel and the use of standard protocols for databases and AI Gateway.
Oct 2025Introducing Vercel Agent: Your new Vercel teammateIntroduces Vercel Agent, an AI teammate for development workflows, available in Public Beta with Code Review and Investigations skills. Code Review operates within GitHub, analyzing diffs, dependencies, and framework patterns, running simulated builds in Vercel Sandbox to validate suggestions before presenting them as pull request feedback. Investigations analyze real-time and historical data to debug production issues, interpreting application context and telemetry to identify root causes and provide actionable next steps. This marks the first product in the Agent-as-a-Service category, offering a production-grade agent that runs securely on behalf of the user, integrated into their workflow. The agentic marketplace is also launched.
Mar 2023Framework-defined infrastructureThis post introduces and defines the concept of Framework-Defined Infrastructure (FdI) as an advancement over Infrastructure as Code (IaC). It explains how FdI leverages framework conventions to automatically provision infrastructure, using Next.js examples (file-based routing, `getServerSideProps`, `getStaticProps`, deferred static generation, middleware, image optimization) to illustrate the mapping of framework concepts to Vercel's infrastructure primitives (serverless functions, static hosting, edge computing). It also contrasts serverless architecture with traditional IaC approaches like Terraform for AWS Lambda and touches upon solving local development challenges for serverless.