Back/How I AI
Blog
How I AI

How I AI: Guillermo Rauch (Vercel CEO) on how to vibe code to production with v0

Vercel CEO Guillermo Rauch joins me to reveal two powerful workflows: how his team uses v0's new Git integration to ship production code for skills.sh, and how he built a creative AI vs. AI chess game from scratch.

Claire Vo's profile picture

Claire Vo

February 4, 20268 min read
How I AI: Guillermo Rauch (Vercel CEO) on how to vibe code to production with v0

In this special live episode of 'How I AI,' I was so excited to sit down with Guillermo Rauch, the CEO of Vercel. We've all seen impressive demos of AI generating user interfaces from a text prompt, but what Guillermo and his team are building with v0 goes so much further. It’s about moving beyond the initial “vibe coding” and into a world where AI is a true collaborator in a production-grade engineering workflow.

Guillermo is at the forefront of this shift, and he gave us a behind-the-scenes look at how Vercel is eating its own dog food. He showed us not just one, but two incredible workflows that showcase the full spectrum of v0’s capabilities. We explored how the Vercel team built and now maintains Skills by Vercel—a viral hub for AI skills—using a complete git-based workflow right inside v0. Then, we looked at a wonderfully creative project, an AI-powered 3D chess game, that shows the platform's power for complex, fun applications.

These examples are more than just cool demos; they represent a fundamental change in how software is made. It’s about lowering the barrier to entry for everyone on the team—designers, marketers, PMs—to contribute directly to the codebase in a safe, reviewable way. It bridges the gap between a great idea and a shipped feature, and I think it’s going to unlock a new level of speed and collaboration for companies everywhere.

Workflow 1: Shipping Production Features with v0 and Git

Our first workflow is a big one. It's about safely iterating on a live, high-traffic application. Guillermo walked us through how the Vercel team iterates on skills.sh, a site that has gone viral with over 34,000 community-submitted skills and hundreds being added every hour. You can't just ship changes to an app like that without a proper process, and Guillermo showed how v0 now bakes that process right in.

Step 1: Branching Directly in v0

The entire process starts with a familiar concept for any engineering team: creating a branch. Instead of using a terminal or a separate Git client, Guillermo showed how you can create a new branch directly from the v0 interface for an existing project. He even followed a common naming convention, g/root-g, which I loved seeing because it makes v0 feel like a true teammate in an established engineering culture.

The Vercel dashboard interface (v0.app) displaying the 'skills-leaderboard' project, complete with its production deployment status, deployed URL (skills-leaderboard-3870vky10.vercel.sh), and an associated 'skills.sh' script. The sidebar shows recent AI-related chat activities such as 'AI chat bubble' and 'Skill analyzer AI'.

Once the branch is created, v0 spins up a complete development environment. This is huge. It has a full VS Code editor, a live preview of the application, and it installs all the necessary dependencies. For anyone less technical, this eliminates the nightmare of setting up a local environment—no more installing homebrew or debugging dependency conflicts.

Step 2: Prompting a New Feature

With the branch ready, it was time to build. The Vercel team had been getting feedback that with so many skills, users needed a way to gauge quality. So, Guillermo prompted v0 to add a rating system. What I found fascinating was how he included production-level thinking directly in the prompt.

"Let's add a five star based rating system for the skills. uh, put it on the sidebar. be mindful that, uh, we should rate limit or prevent abuse on the scores that we receive... and make it make sense within the style of this skills website."
A Vercel v0.app interface demonstrating an AI chat interaction with the prompt 'let's add a 5 star bas', displayed alongside documentation for 'Vercel React Best Practices' and a `npx skills add` command.

v0 understood the context. It knew the project used Redis for its database and generated a rating system that fit the existing "terminal core" aesthetic of the site. The agent handled creating the UI, the backend logic for storing the rating, and considered the security implications mentioned in the prompt.

A preview of the v0 UI for Vercel React Best Practices, demonstrating a generated 5-star rating system and an AI assistant providing development instructions for its creation.

Step 3: Iterating on the Details

After the initial generation, Guillermo noticed a tiny flicker when the page loaded—a layout shift that happens before the rating data is fetched. This is exactly the kind of polish that separates a prototype from a production feature. Instead of opening the code, he just gave v0 another simple instruction:

"Hey, when we don't have data, make sure there's no layout shift"

v0 understood the request and fixed the issue by ensuring the skeleton loader was stable. This iterative loop is incredibly fast and allows you to refine the UI with natural language.

A developer's screen showing an AI chat interaction concerning a layout shift, alongside detailed documentation for 'Vercel React Best Practices' and a terminal command for skill installation. This highlights AI-assisted development workflows for performance optimization.

Step 4: Creating a Pull Request

Once he was satisfied, Guillermo clicked a single button in v0 to create a pull request on GitHub. This is where the workflow really comes together. v0 didn't just push the code; it automatically generated a detailed PR description explaining the changes it made, from adding new UI components to updating the data fetching logic. It’s a huge time-saver and makes code reviews much more efficient.

A Vercel-deployed web application, a 'skills leaderboard' featuring a 'star rating system,' is shown in a browser during an authentication step, highlighted in a podcast video.

Step 5: Reviewing with Preview Deployments

As soon as the PR was created, the Vercel bot kicked in and deployed the branch to a unique preview URL. This isn't just a local dev server; it's a production-like environment running on Vercel's enterprise infrastructure. This allows the whole team to test the new feature in a high-fidelity setting, giving everyone the confidence to merge the change into the main branch. This seamless flow from idea to a production-ready PR is what makes v0 such a powerful tool for real-world development.

A live preview of the skills.sh web application demonstrating 'Vercel React Best Practices,' featuring a 5-star community rating system, weekly install metrics, and a command to add the skill, displayed during a podcast.

Workflow 2: Creative AI Development with the v0 Chess Match

After showing us the serious, production side of v0, Guillermo shared a project that’s pure fun and creativity: an AI chess match he built over the holidays. My kids are obsessed with this one! You can check it out for yourself at v0-chess-match.app. This project demonstrates how you can use v0 to build complex, interactive applications from scratch.

A live demo of the v0 Chess Match app, showcasing a 3D chess board where AI models like Gemini-2.5-flash-lite and gpt-oss-120b are actively playing or analyzing. The interface displays chess notation, model outputs, and real-time thinking processes.

Step 1: From Idea to 3D Render

Guillermo imagined an ESPN-style broadcast of a chess match between two AIs, complete with a 3D board. He was curious if v0 could handle it. He prompted v0 to generate 3D code using Three.js and was able to get a live, interactive chess match running where two AI models battle each other.

Step 2: Streaming AI "Thoughts"

One of the coolest parts of this project is that you can see the AI's reasoning. Guillermo used streaming to display the model's thinking tokens in real-time. You can watch as it considers different moves and explains its strategy. This is powered by the Vercel AI SDK and provides a fascinating glimpse into the model's 'mind.'

It’s a combination of all the Vercel AI infrastructure. It's using a workflow so the game could run forever. The game could literally run forever or until I run out of tokens.

Step 3: Building for Durability

To make sure the game could run indefinitely, Guillermo used Vercel Workflows. This ensures that if a model provider has an outage or a function times out, the system will automatically retry. This is a critical component for building reliable, long-running agentic applications, and it’s something we’ll see v0 help developers create more easily in the future.

Step 4: Debugging with AI

Building something this complex is rarely a straight line. Guillermo ran into an issue where the 3D model he was using had all the chess pieces stuck to the board as a single mesh. To solve this, he created a meta-workflow. First, he asked other AI models to teach him about the 3D file format. Then, he used that knowledge to prompt v0 to build him a custom debug tool inside the app, allowing him to visualize the 3D mesh and textures to solve the problem. This is a perfect example of using AI to help you debug and understand code that another AI wrote.

Democratizing Development

Looking at these two workflows together, you can see the full picture of where AI-assisted development is headed. On one hand, we have a rigorous, safe process for shipping code to production on a high-traffic application. On the other, we have this boundless canvas for creative experiments like the AI chess match.

What I love is that v0 is making both of these accessible to everyone. The friction of going from an idea to a live preview that your team can review is now incredibly low. It changes the dynamic from 'petitioning the engineers' to 'here’s a PR with my idea fully implemented.' This is how you accelerate a company and build a culture of innovation.

I encourage you to try it out. Go to v0.dev, import one of your own GitHub repositories, and see what you can build. The future of software development is collaborative, iterative, and powered by AI.

---

Where to find Guillermo and Claire

Guillermo Rauch

Claire Vo

Tools and References

Try These Workflows

Step-by-step guides extracted from this episode.

Become a 10x PM.
For just $15 / month.

We've made ChatPRD affordable so everyone from engineers to founders to Chief Product Officers can benefit from an AI PM.

How I AI: Guillermo Rauch (Vercel CEO) on how to vibe code to production with v0 | ChatPRD Blog