Rebuilding Tray.ai's Marketing Site on a Modern Stack
Tray.ai, a composable AI integration and automation platform, partnered with Roboto Studio to overhaul its marketing site. The project's goal was to move off a legacy Next.js setup and consolidate siloed infrastructure — content, forms, and code — into a single, unified system. The migration cut build times from a full day down to two minutes.
The old infrastructure was a bottleneck. Running an outdated version of Next.js with Serverless, 250MB package builds, and a rigid dev–stage–prod pipeline made deployments slow and error-prone. Previewing changes meant full deploys, and form logic lived in a separate repository. Build failures were common, and successful ones took an entire day to complete.
Roboto Studio rebuilt Tray.ai's site from the ground up using Next.js. The engineers consolidated everything into a single Turborepo monorepo, implementing custom page resolution and deeply nested dynamic routing to manage the complexity of more than 500,000 pages. The new architecture uses a hybrid rendering strategy, combining Static Site Generation (SSG), Server-Side Rendering (SSR), and Client-Side Rendering (CSR) where each makes the most sense.
Inside the New Architecture
With the migration to Vercel, deployments are now automatic and git-based. Every branch push triggers a build, and developers can review changes using preview deployments that replicate the production environment, catching issues before merging.
Roboto also rebuilt Tray's forms engine using Vercel Functions, integrated with reCAPTCHA for protection. The team monitors real-user interaction through Vercel Analytics and Speed Insights, which provide performance data directly in the Vercel Dashboard. This visibility helps the team identify and address issues immediately rather than digging through layers of abstraction to find broken deploys.
Catching Attacks and Shoring Up Resilience
The improved setup also proved resilient under pressure. Through Firewall Observability, the team detected a DDoS attack targeting dynamic SSR routes. They refactored the most vulnerable pages to static, cached versions, and the problem resolved itself without infrastructure intervention. This flexibility is built into their use of Vercel's Edge Network.
Tray.ai's site now handles over one million page views monthly. The consolidation of multiple repos into one codebase reduced operational complexity, making it easier to manage forms, content, and shared logic in a single location. Combined with edge caching and real-time analytics, the new foundation is designed for faster iteration and scalability.
