Development · Article or course
Pair Programming with gpt-5, Claude, & Cursor (Unfiltered Live Build)
Unfiltered live coding stream building a Next.js AI app with Cursor
A 2.5-hour recorded live build by Ray Fernando and Artem Lazarev constructing a full-stack Next.js AI image editing application from scratch. The video demonstrates real-world AI-assisted engineering using Cursor, the Vercel AI SDK, and OpenAI APIs, comparing model outputs and showing unfiltered live debugging. It is useful for developers wanting to see actual prompt workflows, agent rules, and error resolution rather than an edited tutorial.
What it's for
- Learn practical agentic pair programming workflows using Cursor and LLMs
- Integrate OpenAI image APIs with Next.js and the Vercel AI SDK
- Troubleshoot API and model hallucination issues during live AI code generation
- Set up custom project rule files and duplicate chats in Cursor to evaluate models
Features
- Complete build of a full-stack Next.js image editing app from an empty directory
- Side-by-side comparison of model planning and UI scaffolding
- Demonstration of Cursor features including agent rule generation and duplicate chats
- Integration walkthrough of the Vercel AI SDK v5 with OpenAI image editing APIs
- Detailed chapter timestamps covering setup, prompting, debugging, and final deployment tips
Pricing
Free
Details
- Type
- Article or course
- Category
- Development
- Topic
- AI Coding Tutorials
- Open source
- No
- Platforms
- Web
- For
- Software developers and AI engineers building web apps with Cursor and modern LLMs
- Tags
- cursor, nextjs, openai, claude, live coding, pair programming, ai development
Why it was saved
- @RayFernando1337: A watchable, unfiltered session showing a real workflow combining GPT-5 and Claude Code with Cursor.
Related
An 8-step framework for building apps with AI using structured documentation
Install AI rule files into projects using shadcn CLI registries
System prompt to refactor web apps for 100k+ programmatic SEO pages
Copy-paste React UI components built with Tailwind CSS and Radix primitives.
Pre-built, responsive shadcn/ui blocks and layouts for React applications
Portable agent skills that enforce design quality in AI-generated frontends
Updated