# Mantua Protocol Newsletter

*From Vibe Coding to UX Foundation*

By [Mantua.AI](https://paragraph.com/@mantua-ai) · 2025-08-31

---

### **Intro**

Mantua Protocol started, like most things do, with vibes.

During the first hackathon, I hacked together the MVP in **Replit**. It looked good enough to demo, but under the hood it was clunky. Only two-thirds of the app worked. The **swap flow worked. The add liquidity flow worked.** But the most important part of the vision — the ability to **chat with the large language model** — didn’t.

That’s the reality of “vibe coding”: it gets you to _something,_ but rarely all the way to _working._

![](https://storage.googleapis.com/papyrus_images/5f4b3834f959f3f254956b52701ec639.png)

### **Phase 1 → The Limits of Vibe Coding**

Vibe coding was a double-edged sword. It gave speed, but it also created cracks:

*   The LLM chat interface, which was the heart of Mantua, didn’t function.
    
*   Smart contracts can’t be vibe-coded — they require precision.
    
*   Data must be accurate; hallucinations aren’t acceptable in DeFi.
    

We had enough to submit. Enough to demo. But not enough to believe this was a product that could scale.

### **Phase 2 → Structured Prototyping in Replit**

The second version pushed things further. This time, the chat interface _did_ work as designed. The swap and liquidity flows were tighter. The UI was lighter. For the first time, Mantua felt like a DeFi product instead of a demo.

![](https://storage.googleapis.com/papyrus_images/c0ccf9e1b8a16620b20dae2d4e53f9bf.png)

But the challenges of building a DeFi product didn’t go away:

*   Vibe coding can’t ship reliable smart contracts.
    
*   LLMs need structured prompts and accurate datasets.
    
*   Every small detail in UX matters when real value is moving.
    

So while the product looked better, building it was still a grind.

### **Phase 3 → UX Foundation with Onchain Design Sprint**

The breakthrough came when Mantua was selected for **FBI Africa’s Onchain Design Sprint** — only 26 projects chosen out of 100+ applicants. We worked with a dedicated design team to audit and redesign Mantua’s UX from the ground up.

![](https://storage.googleapis.com/papyrus_images/6160ccbd113cb4493768ca4f472e38c8.png)

This was where Mantua grew up.

*   We got an **audited UX** that removed the friction points, pushing users away.
    
*   A **design system** that was clean, scalable, and professional.
    
*   A product experience that matched the ambition: making programmable liquidity conversational.
    

![](https://storage.googleapis.com/papyrus_images/0d8b9d5153916828c6cc2b416a8600c1.png)

### **Why It Matters**

The leap from **vibe coding → structured prototyping → design sprint foundation** wasn’t just cosmetic. It was existential.

Great tech doesn’t matter if users can’t use it. Adoption lives in the UX.

Now, Mantua has that UX foundation. Not vibes. Not clunky half-working demos. A real design system we can build on top of — one that guides, simplifies, and converts.

Mantua’s journey so far has been about evolving from spark → prototype → foundation. The next phase is building on top of that foundation and bringing programmable liquidity fully on-chain.

👉 Subscribe [https://landingmantuaprotocol.com/](https://landingmantuaprotocol.com/) to follow along as we turn Mantua’s design system into a live product

---

*Originally published on [Mantua.AI](https://paragraph.com/@mantua-ai/mantua-protocol-newsletter)*