Vibe Coding Works — Here’s the Plugin That Proves It

After several attempts filled with mistakes — where I was too focused on building a plugin as quickly as possible — I realized that rushing doesn’t work. So I decided to step back and go deep into the fundamentals. I spent time learning how WordPress plugins are structured, what languages are typically used, and what it really takes to build a reliable plugin like this one.

Once I had a solid understanding, I shifted into a more intuitive and creative approach — what I like to call vibe coding. That’s when the idea behind Suggester finally came to life.

What Is Suggester?

Suggester is a WordPress plugin that helps you create smart, AI-powered suggestion tools. It lets you display real-time content suggestions to your visitors based on the keywords they type — by connecting to external AI services like Google Gemini and OpenRouter.

🟢 You can find it on the WordPress plugin directory by searching Suggester 🔗 Or download it directly here: wordpress.org/plugins/suggester 📂 Open-source on GitHub: github.com/webcava/suggester 🌐 Official website: webcava.com

Tools I Used to Develop the Plugin

While developing the Suggester plugin, I relied on tools that helped me overcome obstacles and speed up the building process. The first tool was ChatGPT, where I asked all my questions: What do I need? What programming languages are suitable? Any issue or confusion I ran into — even if it seemed trivial — I would take it straight to the chat and get clear answers that helped me move forward.

The second tool was Cursor AI, which I used to write almost the entire plugin through simple text prompts. At the beginning, I prepared a file like prd.md where I explained the idea in detail: what the plugin is, which languages I wanted to use, how the tool should look, and other essential details. After that, I started working with Cursor as if it were my teammate on the project.

At that time, I was using the Sonnet 3.7 model within Cursor, and it performed really well, helping me complete tasks efficiently. Sonnet 4 hadn’t been released yet during the development of the plugin, but once it came out and I tried it, I noticed a clear improvement in professionalism and output quality — making it an even stronger tool for anyone looking to build similar projects.

What Can You Do with Suggester?

• Build up to 3 custom AI suggestion tools
• Use simple shortcodes to embed them anywhere on your site
• Customize prompts and output formatting to fit your needs
• Track usage with built-in stats
• Easily manage multiple API keys and AI models

How Does It Work?

1. You create a tool by setting up a custom prompt with your preferred AI model
2. Embed the tool on your site using a shortcode
3. Visitors enter a keyword or phrase
4. The plugin sends that input to the selected AI service
5. The AI returns suggestions instantly, which are shown on the page

Privacy & Data Use

Suggester does not collect or store any personal data. Only the keyword entered by the visitor and the prompt you configured are sent to the AI service (Google Gemini or OpenRouter).

When You Know the rules, You Can Play Freely

Suggester came to life when I stopped rushing and started coding with intention. You don’t need to be a pro developer — but you do need to understand the rules. Once you do, vibe coding can take you further than you expect.