Building Full Stack AI Apps with eve & Next.js

One Repo, Two Runtimes: Build and Deploy a Durable AI Agent from Scratch

Build Full Stack AI Applications with eve and Next.js

What happens when a Next.js application and an AI agent share the same codebase? In this hands-on tutorial, you’ll build a complete full stack pizza delivery application from scratch, combining Next.js, eve, SQLite, and Drizzle in one repository. You’ll learn how two runtimes can work side by side while building a real application where an AI agent can understand orders, interact with data, and drive application workflows.

Build Full Stack AI Applications with eve and Next.js

From AI Agent to Production Ready Application

Go beyond simple model calls and build the architecture that makes an agent useful in a real application. You’ll work with typed tools, durable state, human in the loop approval, skills, and evals, while seeing exactly where agent logic meets traditional application code. By the end, you’ll have a working full stack AI application deployed to Vercel and a practical understanding of what eve’s file system first approach brings to agentic development.

From AI Agent to Production Ready Application

Lesson-Overview

Lesson 1: Two Runtimes, One Repo: Scaffolding the Storefront

Start from an empty folder and build the foundation for a Next.js app and eve agent in one repository. You’ll connect both runtimes with withEve(), set up SQLite and Drizzle for the menu data, and wire the first agent interaction into the storefront with useEveAgent(). By the end, both runtimes work together as a single project with files as their interface.

Turn the agent into an active part of the application with typed tools, Zod schemas, and Drizzle. You’ll build the custom pizza pricing flow and use defineState to maintain durable order state across conversations, including size, crust, toppings, and price. See how agent logic and application code work together to create a real time ordering experience.

Add human in the loop approval to ensure orders are only placed after explicit customer confirmation. Then build the kitchen workflow with order status tools and real time updates that move orders from prep to bake to delivery. Learn how eve’s durable session model supports reliable, multi step application workflows.

Take the application from working prototype to production ready reference app. Add on demand skills for intelligent topping recommendations, use defineEval to test the complete ordering flow, and catch regressions before deployment. Finally, deploy both runtimes to Vercel and see how the file system first approach translates from local development to production.

Expert Knowledge for...

  • Full Stack Developers, who want to combine Next.js with agentic application development.

  • AI Developers, who want to build agents with tools, state, skills, and human approval.

  • Next.js Developers, who want to integrate durable AI agents into real applications.

  • Software Architects, who want to understand how agent logic and application code work together.

Expert Knowledge for
Complete this Tutorial and Learn how to...

Complete this Tutorial and Learn how to...

  • build a full stack AI application with Next.js and eve.
  • connect two runtimes in one repository with withEve().
  • create agents with tools, durable state, and human approval.
  • implement human in the loop workflows, skills, and agent evals.
  • deploy a complete agentic application to Vercel.

Expert-Led Knowledge for Real-World Applications

Nir Kaufman - Tikal

Nir Kaufman is a software architect, consultant, and international speaker specializing in modern web development, TypeScript, Angular, and AI-powered applications. He helps development teams build scalable frontend architectures and modern full-stack systems, with a strong focus on developer experience, performance, and emerging AI workflows. Nir is known for his practical, hands-on teaching style and regularly shares his expertise through workshops, conference talks, and technical training sessions.

Nir Kaufman

Register Now and Complete This Tutorial

Fullstack Experience

Individual Membership
$ 10
00
Monthly
  • $100 off 3 conference tickets every year
  • Access all Live Events, Read content & Courses
  • 6 month access to conference recordings

Fullstack Elevate

Corporate Membership from 5 users
Inquire Now
  • Access all content on the platform
  • Up to 28% off conference tickets
  • Book your tickets directly on the platform
  • Training insights for team leads
  • easy training approval system

Already have Fullstack?

The tutorial is available on demand. Learn at your own pace and revisit the lessons whenever you need. Want to complete more Fullstack Tutorials? Browse through the complete list here.

Gen AI Engineering Days 2024

Live on 29 & 30. October 2024 | 13:00 – 16:30 CEST