We’ve all been there: mindlessly scrolling through the internet when we stumble upon something we absolutely must have. For me, it was a video of a digital pet cat living on a desktop. The animations were incredibly cute, and I loved the idea of having a tiny companion on my screen.
There was just one problem—it cost money.
So, I did what any developer would do: I thought, "Why not just build one myself?"
Bringing the Pet to Life
I started working on it. Let me tell you, getting the animations right was not an easy task, but after a lot of trial, error, and tweaking, I managed to get a solid set of interactions working.
My digital pet could finally react to me:
- Patting makes the pet blush.
- Idle time causes the pet to fall asleep.
- Typing makes it type right alongside you.
- Movement makes it follow your cursor across the screen.
- Double-tapping makes it jump in joy.
It was fun, it was cute, but then a realization hit me. It was just another pet engine sitting on my desktop, eating up RAM and CPU usage for absolutely no reason other than looking adorable.
I wanted it to be more than just a screen decoration.
The Pivot: Giving Paendeo a Job
After brainstorming and getting some great suggestions, I decided to shift gears. I moved the project locally into a web extension and decided to give my pet an actual purpose.
I named it Paendeo (Korean for Panda).
In today’s era of AI, almost everyone relies on tools like Gemini, ChatGPT, Claude, or Perplexity. However, getting a great response out of an LLM requires writing a great prompt—something that establishes the right context, role, and constraints. Not everyone has the time or knowledge to prompt-engineer every single query.
That’s where Paendeo comes in.
The animations stay the same, but now, Paendeo works in the background as a prompt optimizer. I built a pipeline where the pet analyzes the text you write in the AI's chatbox. If your prompt isn't optimized, Paendeo automatically enhances it by:
- Assigning the correct role to the model.
- Structuring the context you inputted.
- Adding necessary constraints for a cleaner output.
The Difference it Makes: If you type a lame prompt like: "Write an essay about money usage and responsibility," you'll likely get a generic, boring output. Paendeo intercepts this and rewrites it into a near-perfect prompt, guaranteeing a response that is multiple times better and highly tailored to your needs.
Under the Hood
If you're curious about the tech stack, the whole project is built using:
- Frameworks & Runtimes: Electron, Node.js
- Languages: JavaScript (ES6+) for the core logic, Python for asset processing
What’s Next? (I need your help!)
While I am incredibly proud of how Paendeo turned out, I am currently at a crossroads. It started as a random desire for a cute desktop pet and evolved into a genuinely useful productivity tool.
Now, I want to take it further, and I am looking for a community of developers, designers, and AI enthusiasts to help me figure out the next steps. I am aiming to release this as an open-source extension for everyone to download, but I need help brainstorming ideas to improve both the web extension and potentially a standalone desktop application.
I would love to hear your thoughts, feedback, and constructive criticism! If you find the project interesting, I would hugely appreciate your support and contributions on GitHub.
Check out the repository (and drop a star if you like it!): Paendeo on GitHub
Let's build something awesome together. Drop your feedback below!
