← Back to Blog
Web3 min read

Web Development 2026

David Broadhurst

Putting together a definitive web development stack in 2026 is harder than it sounds. Everything is shifting fast, especially since Artificial Intelligence (AI) entered the picture. What worked two years ago might already feel outdated. Here's how my current production stack breaks down, and why I landed on each piece.

Editor & AI

Visual Studio Code remains my editor of choice. There's a growing movement toward Command Line Interface (CLI)-driven AI workflows, where you pipe commands straight to a Large Language Model (LLM) to generate code autonomously. I've tried it, but I still prefer keeping eyes on the changes as they happen and making sure each commit actually makes sense.

For AI assistance, Claude Sonnet 4.6 is my go-to LLM for programming. It consistently comes up with better solutions than the alternatives, and features like skills give you additional control over how it behaves. I recently switched from GitHub Copilot to using Claude directly inside VS Code via the Claude app, and it's been a clear improvement.

Hosting & Frontend

Vercel has been my hosting platform for a long time. It simplifies the entire pipeline — from Continuous Integration/Continuous Deployment (CI/CD) through to production hosting. Pair that with React as the component library and you have a solid, reliable front end setup.

Analytics & Marketing

Once an app is live, you need to understand how people are actually using it. Google Analytics is the right call for production apps since it ties in neatly with other Google services. I've also spent a fair amount of time with Google Ads — I was skeptical for a long time, but a well-run campaign really can drive sales. Google Tag Assistant is essential for making sure Ads is picking up the right signals, and Google Search Console is invaluable for understanding how Google is indexing your site.

Auth & Error Tracking

I've tried just about every auth solution going. Right now Clerk is the one I keep coming back to. It has most of the features you'd expect, but the recent addition of Stripe billing support makes it stand out from the crowd. For error tracking, Sentry is a necessity — bugs happen, and you need to know about them before your users do.

Stack Summary

Tool Purpose
VS Code Editor
Claude Sonnet 4.6 AI assistant (LLM)
Claude App AI interface in VS Code
Vercel Hosting & CI/CD
React Frontend framework
Google Analytics Usage analytics
Google Ads Marketing
Google Tag Assistant Tag verification
Google Search Console SEO monitoring
Clerk Authentication & billing
Sentry Error tracking

Three Tools Where I've Made a Clear Choice

If you're building something similar and weighing up your options, here's where I've landed:

Editor: VS Code vs Cursor vs JetBrainsVS Code wins for me. Cursor is interesting but adds complexity I don't need. JetBrains is solid for Java-heavy work but overkill for this kind of stack.

AI Model: Claude Sonnet 4.6 vs ChatGPT-4o vs GeminiClaude every time. Better reasoning for code, more predictable output, and the skills feature gives you a level of control the others don't match yet.

Authentication: Clerk vs Auth0 vs NextAuthClerk is the pick. Auth0 gets expensive fast once you scale. NextAuth requires a lot more legwork to set up properly. Clerk lands right in the sweet spot.

— David