Free Vibe Coding Practice: AI Blogger Agent Open Source Release

We are releasing a free Vibe Coding practice program as open source. It guides you through creating an AI Blogger Agent with Caret, allowing anyone to experience Vibe Coding hands-on.

Recommended for

• Those curious about "What exactly is an AI agent?"
• Those who can build prototypes but get stuck with practical application.
• Those concerned about token costs and keeping up with the latest tech trends.

When information is abundant but a clear path is hard to find, hands-on practice can be the quickest answer. Let Caret help you clear the fog.

Is AI creation/development all just a Vibe?

To alleviate some of this uncertainty and curiosity, and to help you take a real first step in AI development, we are today unveiling a special gift we've been preparing for a long time. 🎁 It's an open-source educational program where you can learn 'real Vibe Coding' with Caret, the AI coding tool from Caretive, and experience the entire process of developing an AI Blogger Agent. (For those who are curious, you can check out the full source code and educational materials on our GitHub repository right now!)

multi-post-agent/README.md at main · aicoding-caret/multi-post-agent

This project is an educational project aimed at developing an AI-based multilingual blog posting automation agent. It is conducted in a "Vibe Coding" style with the Caret AI assistant and provides hands-on experience in AI agent development and LangGraph utilization. - aicoding-caret/multi-post-agent

github.com

This program offers free Vibe Coding practice and provides warm guidance and a solid practice environment to help you overcome your difficulties and create a working product beyond just a prototype.

​

In this article, I'll whisper the details of how this new educational program can be a cozy companion on your AI development journey, and why 'Vibe Coding' with Caret is such a special, heart-pounding experience. I sincerely hope this program and article will be a small beacon for those navigating the fog of AI development. 🌟

1. AI Agents: Let's Build and Learn Together Now! 📖

"What is an AI agent?" As they say, seeing is believing, right? Through this program, you will get to understand what an AI agent is by building one yourself. You'll be able to create a simple yet powerful AI agent that transforms user-input text into a desired style and automatically posts it to a specified Blogger.com blog. Isn't that exciting? 😊

"How can I build it?" Don't worry too much! With great open-source tools like LangGraph and FastAPI, and most importantly, your trusty AI development companion Caret (and a little help from me, Alpha! 😉), it's totally possible. The docs/education-scenario.md file provided in this program will warmly guide you through your systematic learning journey, from setting up the development environment to completing the AI agent step-by-step.

​

This educational program is not just a code sample. It's a living learning process where you can experience AI agent development from A to Z, as if creating a story. Through the process of tackling challenges, building, and learning, AI agents will no longer feel like a difficult and distant concept!

2. The Secret to 'Real Vibe Coding'? Learn with This Program:

Friendly Communication with AI + Meticulous Records! ✍️🤝

"The prototype looks plausible, but why can't I get any further?" Many people are excited about the infinite potential of AI but struggle to create service-level results. Why is that? Perhaps because AI is not a magic wand. 🪄 What's truly important is the 'tiki-taka' with AI—clear communication that shares intent, 'recording' those precious processes, and having the AI 'reference' those 'records' to grow together.

This illustrates a systematic documentation and development process with AI. It captures the core of document-based AI collaboration, starting from requirement specifications to system architecture, task lists, and documentation guides, creating AI instructions as a .caretrules file, which the AI then repeatedly uses as input. The philosophy that "documentation is a contract with AI, not a record," and "rules are a sustainable context, not a command" is crucial.

​

In this educational program, you will learn and practice effective collaboration with Caret AI in the following ways:

​

No more "Just figure it out!" 🙅‍♀️: Instead of giving vague instructions to the AI, you will learn to create and utilize clear blueprints like README.md, docs/requirement.md (requirement specification), and docs/development-guide.md (development guide) together with the AI. These documents are a precious promise with the AI about what we want to build and how, and they serve as the foundation for warm communication.

Setting AI Behavior Guidelines 📜: You will learn how to set custom behavior guidelines for Caret AI using the .caretrules file in the project root. This allows you to gently guide the AI to accurately understand the project's context and perform tasks consistently.

Specific Task Instructions and Progress Tracking 📊: You'll learn how to give specific tasks to the AI using individual task documents in the docs/tasks/ folder, and how to manage the process by having Caret write code, update documents, and even report task completion (updating checklists, recording logs). It's like keeping a journal together!

This shows how tasks are managed in a real project (In Progress, Pending, Done lists) and a specific work log example using work-logs. Collaboration with AI is not just a simple conversation but can be built on a systematic project management system.

Successful "Vibe Coding" is not about blindly trusting the AI's abilities, but a process where humans and AI clearly define their roles and build together, communicating through the common language of documentation. This program will give you a little peek into the warm secret of moving beyond prototypes to create services that actually work.

​

3. Learn AI Development Without Worrying About Token Costs! Smartly with Open Source and Free Tiers! 💰➡️🆓

I've noticed that many people hesitate to get into AI development, especially using LLMs, because of the "token costs." However, this educational program guides you to choose smart tools so you can enjoy learning AI development without the financial burden:

Google Gemini API: The Gemini API, currently available through Google AI Studio, offers a generous free usage tier, making it stress-free for personal learning or small projects. (As of June 2025, we are configuring the free learning environment based on the Gemini 2.0 Flash or 2.5 Pro models. Isn't that great? 😊)

Hooray for Open Source! 🙌: The core tech stack of the project, including FastAPI and LangGraph, are all powerful and kind-hearted open-source tools.

Caret AI: And our trusty development companion, Caret, can currently be installed for free via a VSIX file. With just a Gemini API key, you can start using it right away. (Don't worry, you can also see the estimated cost per conversation within the Caret interface!)

Free Development Environment Setup for the Multi Post Agent Open Source Project

From the Caret-powered Multi Post Agent open source education slide (Multi Post Agent open source project by Caretive)

​

By using these tools smartly, you can learn and experiment with AI development without worrying about the costs. This educational program will be a great friend on your burden-free AI development learning journey! ​

4. Conclusion: A Lighthouse for Developers Navigating the AI Era, and the Bright Future We'll Build Together! 🌟

This educational program and open-source project go beyond simple technology transfer to present a new perspective and a warm collaborative method for developers in the AI era. I believe we are not heading towards an "era without developers," but rather, we should prepare for an era of developers who become more powerful and can achieve more with AI.

To the junior developers feeling anxious, senior developers contemplating new challenges, and non-developer CEOs considering AI adoption, I hope this program offers a little comfort and a concrete starting point. "Vibe Coding" is not just a tech trend. It's the heart-pounding future of development culture, where we grow and create new value with AI. 💕

​

So, why release all of this for free and as open source?

Because we are eagerly awaiting your sparkling interest in Caret AI and the new development method called "Vibe Coding"! 🔥

This program and the accompanying open-source project (GitHub - aicoding-caret/multi-post-agent) are like a tiny seed just beginning to sprout. We sincerely hope to contribute to the open source together, and to contemplate and build better AI collaborative development methodologies. Your precious ideas and warm participation can make Caret blossom beautifully. 🌸

And to be honest... (wink 😉) we at Caret are planning to launch a commercial service with an even more amazing look soon! We hope you'll get a preview of Caret's infinite possibilities through this educational program and show a lot of love and interest for our upcoming service! 🙏

Jump into the warm world of "Vibe Coding" right now and let's create a brilliant future together with Caret!