Vibe coding is a new way to create software using AI tools such as ChatGPT, Cursor, Replit, and Gemini. It works by describing to the tool what you want in plain language and receiving written code in return. You can then simply paste the code into an environment (such as Google Colab), run it, and test the results, all without ever actually programming a single line of code.

Collins Dictionary named “vibe coding” word of the year in 2025, defining it as “the use of artificial intelligence prompted by natural language to write computer code.”

In this guide, you’ll understand how to start vibe coding, learn its limitations and risks, and see examples of great tools created by SEOs to inspire you to vibe code your own projects.

Vibe coding variations

While “vibe coding” is used as an umbrella term, there are subsets of coding with support or AI, including the following:

Type	Description	Tools
**AI-assisted coding **	AI helps write, refactor, explain, or debug code. Used by actual developers or engineers to support their complex work.	GitHub Copilot, Cursor, Claude, Google AI Studio
Vibe coding	Platforms that handle everything except the prompt/idea. AI does most of the work.	ChatGPT, Replit, Gemini, Google AI Studio
No-code platforms	Platforms that handle everything you ask (“drag and drop” visual updates while the code happens in the background). They tend to use AI but existed long before AI became mainstream.	Notion, Zapier, Wix

We’ll focus exclusively on vibe coding in this guide.

With vibe coding, while there’s a bit of manual work to be done, the barrier is still low — you basically need a ChatGPT account (free or paid) and access to a Google account (free). Depending on your use case, you might also need access to APIs or SEO tools subscriptions such as Semrush or Screaming Frog.

Your customers search everywhere. Make sure your brand shows up. The SEO toolkit you know, plus the AI visibility data you need. Start Free Trial Get started with

To set expectations, by the end of this guide, you’ll know how to run a small program on the cloud. If you expect to build a SaaS or software to sell, AI-assisted coding is a more reasonable option to take, which will involve costs and deeper coding knowledge.

Vibe coding use cases

Vibe coding is great when you’re trying to find outcomes for specific buckets of data, such as finding related links, adding pre-selected tags to articles, or doing something fun where the outcome doesn’t need to be exact.

For example, I’ve built an app to create a daily drawing for my daughter. I type a phrase about something that she told me about her day (e.g., “I had carrot cake at daycare”). The app has some examples of drawing styles I like and some pictures of her. The outputs (drawings) are the final work as they come from AI.

When I ask for specific changes, however, the program tends to worsen and redraw things I didn’t ask for. I once asked to remove a mustache and it recolored the image instead.

If my daughter were a client who’d scrutinize the output and require very specific changes, I’d need someone who knows Photoshop or similar tools to make specific improvements. In this case, though, the results are good enough.

Building commercial applications solely on vibe coding may require a company to hire vibe coding cleaners. However, for a demo, MVP (minimum viable product), or internal applications, vibe coding can be a useful, effective shortcut.

How to create your SEO tools with vibe coding

Using vibe coding to create your own SEO tools require three steps:

1. Write a prompt describing your code
2. Paste the code into a tool such as Google Colab
3. Run the code and analyze the results

Here’s a prompt example for a tool I built to map related links at scale. After crawling a website using Screaming Frog and extracting vector embeddings (using the crawler’s integration with OpenAI), I vibe coded a tool that would compare the topical distance between the vectors in each URL.

This is exactly what I wrote on ChatGPT:

I need a Google Colab code that will use OpenAI to:

*Check the vector embeddings existing in column C. Use cosine similarity to match with two suggestions from each locale (locale identified in Column A). *

The goal is to find which pages from each locale are the most similar to each other, so we can add hreflang between these pages.

I’ll upload a CSV with these columns and expect a CSV in return with the answers.

Then I pasted the code that ChatGPT created on Google Colab, a free Jupyter Notebook environment that allows users to write and execute Python code in a web browser. It’s important to run your program by clicking on “Run all” in Google Colab to test if the output does what you expected.

This is how the process works on paper. Like everything in AI, it may look perfect, but it’s not always functioning exactly how you want it.

You’ll likely encounter issues along the way — luckily, they’re simple to troubleshoot.

First, be explicit about the platform you’re using in your prompt. If it’s Google Colab, say the code is for Google Colab.

You might still end up with code that requires packages that aren’t installed. In this case, just paste the error into ChatGPT and it’ll likely regenerate the code or find an alternative. You don’t even need to know what the package is, just show the error and use the new code. Alternatively, you can ask Gemini directly in your Google Colab to fix the issue and update your code directly.

AI tends to be very confident about anything and could return completely made-up outputs. One time I forgot to say the source data would come from a CSV file, so it simply created fake URLs, traffic, and graphs. Always check and recheck the output because “it looks good” can sometimes be wrong.

If you’re connecting to an API, especially a paid API (e.g., from Semrush, OpenAI, Google Cloud, or other tools), you’ll need to request your own API key and keep in mind usage costs.

Should you want an even lower execution barrier than Google Colab, you can try using Replit.

Simply prompt your request and the software will create the code, design, and allow testing all on the same screen. This means a lower chance of coding errors, no copy and paste, and a URL you can share right away with anyone to see your project built with a nice design. (You should still check for poor outputs and iterate with prompts until your final app is built.)

Keep in mind that while Google Colab is free (you’ll only spend if you use API keys), Replit charges a monthly subscription and per-usage fee on APIs. So the more you use an app, the more expensive it gets.

Inspiring examples of SEO vibe-coded tools

While Google Colab is the most basic (and easy) way to vibe code a small program, some SEOs are taking vibe coding even further by creating programs that are turned into Chrome extensions, Google Sheets automation, and even browser games.

The goal behind highlighting these tools is not only to showcase great work by the community, but also to inspire, build, and adapt to your specific needs. Do you wish any of these tools had different features? Perhaps you can build them for yourself — or for the world.

GBP Reviews Sentiment Analyzer (Celeste Gonzalez)

After vibe coding some SEO tools on Google Colab, Celeste Gonzalez, Director of SEO Testing at RicketyRoo Inc, took her vibing skills a step further and created a Chrome extension. “I realized that I don’t need to build something big, just something useful,” she explained.

Her browser extension, the GBP Reviews Sentiment Analyzer, summarizes sentiment analysis for reviews over the last 30 days and review velocity. It also allows the information to be exported into a CSV. The extension works on Google Maps and Google Business Profile pages.

Instead of ChatGPT, Celeste used a combination of Claude (to create high-quality prompts) and Cursor (to paste the created prompts and generate the code).

AI tools used: Claude (Sunner 4.5 model) and Cursor

APIs used: Google Business Profile API (free)

Platform hosting: Chrome Extension

Knowledge Panel Tracker (Gus Pelogia)

I became obsessed with the Knowledge Graph in 2022, when I learned how to create and manage my own knowledge panel. Since then, I found out that Google has a Knowledge Graph Search API that allows you to check the confidence score for any entity.

This vibe-coded tool checks the score for your entities daily (or at any frequency you want) and returns it in a sheet. You can track multiple entities at once and just add new ones to the list at any time.

The Knowledge Panel Tracker runs completely on Google Sheets, and the Knowledge Graph Search API is free to use. This guide shows how to create and run it in your own Google account, or you can see the spreadsheet here and just update the API key under Extensions > App Scripts.

AI models used: ChatGPT 5.1

APIs used: Google Knowledge Graph API (free)

Platform hosting: Google Sheets

Inbox Hero Game (Vince Nero)

How about vibe coding a link building asset? That’s what Vince Nero from BuzzStream did when creating the Inbox Hero Game. It requires you to use your keyboard to accept or reject a pitch within seconds. The game is over if you accept too many bad pitches.

Inbox Hero Game is certainly more complex than running a piece of code on Google Colab, and it took Vince about 20 hours to build it all from scratch. “I learned you have to build things in pieces. Design the guy first, then the backgrounds, then one aspect of the game mechanics, etc.,” he said.

The game was coded in HTML, CSS, and JavaScript. “I uploaded the files to GitHub to make it work. ChatGPT walked me through everything,” Vince explained.

According to him, the longer the prompt continued, the less effective ChatGPT became, “to the point where [he’d] have to restart in a new chat.”

This issue was one of the hardest and most frustrating parts of creating the game. Vince would add a new feature (e.g., score), and ChatGPT would “guarantee” it found the error, update the file, but still return with the same error.

See the complete picture of your search visibility. Track, optimize, and win in Google and AI search from one platform. Start Free Trial Get started with

In the end, Inbox Hero Game is a fun game that demonstrates it’s possible to create a simple game without coding knowledge, yet taking steps to perfect it would be more feasible with a developer.

AI models used: ChatGPT

APIs used: None

**Platform hosting: **Webpage

Vibe coding with intent

Vibe coding won’t replace developers, and it shouldn’t. But as these examples show, it can responsibly unlock new ways for SEOs to prototype ideas, automate repetitive tasks, and explore creative experiments without heavy technical lift.

The key is realism: Use vibe coding where precision isn’t mission-critical, validate outputs carefully, and understand when a project has outgrown “good enough” and needs additional resources and human intervention.

When approached thoughtfully, vibe coding becomes less about shipping perfect software and more about expanding what’s possible — faster testing, sharper insights, and more room for experimentation. Whether you’re building an internal tool, a proof of concept, or a fun SEO side project, the best results come from pairing curiosity with restraint.

Contributing authors are invited to create content for Search Engine Land and are chosen for their expertise and contribution to the search community. Our contributors work under the oversight of the editorial staff and contributions are checked for quality and relevance to our readers. Search Engine Land is owned by Semrush. Contributor was not asked to make any direct or indirect mentions of Semrush. The opinions they express are their own.