From Tokens to Apps — The Next Major Narrative

Pumpfun gave us hundreds of thousands of new tokens

Vibecoding will do the same for web3 apps

$GOAT was the first runner in the AI Agents.

It successfully combined two hot narratives at the time: AI Agents + Memecoins, making the “token value” instantly clear for degens.

However, GOAT didn’t come out of nowhere, It was preceded by a series of major Web2 catalysts — infra by X and increasing momentum from OpenAI.

If you were active in the market back then, you probably remember what happened next.

Shortly after GOAT, everyone started talking about Virtual Protocol, a project that had existed long before the narrative took off and had been trading around $20M–$50M FDV for almost a year.

A very similar pattern is forming right now around vibecoding, a term that only appeared recently. Some degens were already vibecoding simple on-chain projects, but due to the lack of infra, especially in web3, these apps never received enough attention.

Below, I’ll break down:

• The rise of vibecoding in Web2
• its inevitable impact on Web3

The early infra forming on-chain

Bullish signals from Web2

From May 2025 to December 2025, the share of all LLM queries related to programming grew from 11% to 50%.

In less than a year, coding became half of all LLM usage across every category.

At the same time, the valuation of the most popular AI-powered dev platform increased 72x, from $400M to $29.3B in just one year.

If we zoom out and look at survey data and research, there are three more important signals worth highlighting:

• 84% of respondents are using or planning to use AI tools in their development process
• 41% of all code is now AI-generated
• Developers report that AI tools increase productivity by 10–30% on average

It’s also important to clarify that vibecoding is not the same as AI-assisted development.

For complex production-grade systems, relying purely on vibecoding can be risky. That said, given how fast this space is scaling, more devs will start using vibecoding as part of their workflow.

(According to the Stack Overflow Developer Survey 2025, only 17% of developers are vibecoding, which leaves massive room for growth)

The PumpFun effect

Just a few years ago, launching your own token wasn’t easy.

Teams experimented with different launch mechanics until pumpfun introduced a dead simple model.

Pf gave simplicity to creators and a new trading experience to degens.

So why wouldn’t we see the same playbook, but for web3 apps?

Degens are always the first users and early adopters of emerging narratives.

Many of them deeply understand what products the market needs rn, how to launch them, and what CT actually cares about — simply because they’ve been watching new launches for years.

Giving degens the ability to turn ideas into products without coding skills or high costs is essentially giving the trenches fresh ideas to play with.

As a result, we’re likely to see hundreds of new creative experiments across multiple verticals:

Consumer apps, trading terminals, SocialFi, prediction markets, GameFi ponzis, and more.

This, in turn, could accelerate growth in specific narratives of the market.

Beyond the trenches: AI inside major protocols

Beyond the abstract “trenches narrative,” there’s another direction that excites me. AI integration directly into large protocols.

The simplest example is trading terminals.

Delta-neutral strategies, points farming, arbitrage — all of this can be turned into a personalized bot or feature, built by the user directly on the platform using a simple prompt.

For example:

Monitoring funding rates for tokenized stocks and automatically opening positions on two perps DEXes when a significant spread appears. In practice, this approach allows platforms to cover thousands of small user needs that teams often don’t even realize exist, despite being relatively simple to implement.

“We already have Cursor, Replit, Lovable — why web3 forks?”

At some point, I asked myself the same question.

After digging through a lot of material, the answer became pretty clear.

Even within the web2 vibecoding space, there is no single winner. Each platform is unique, with its own focus, features, and target audience.

Some products solve 1-2 specific problems extremely well.

Others try to do everything and end up doing nothing right.

Web3 vibecoding platforms won’t suffer from competition with Lovable or Replit.

Their infra is built specifically for on-chain ideas, not for web2 apps

Web3 Vibe Ecosystem

As of December 2025, there are still very few projects focused on vibecoding or dev tools in web3 and even fewer successful cases where non-devs shipped meaningful web3 products.

Still, I think it’s important to walk through the current landscape.

Apps/Platforms

@ dreamspaceHQ — Without any coding knowledge, you can build websites, platforms, and dashboards.

Dreamspace’s key advantage is that, in addition to creating or connecting smart contracts via prompts, you can also aggregate data across multiple chains:

BTC, ETH, zkSync, Sei, Base and even Stellar wtf?

@ Minidevx — A vibecoding platform on Base, focused on building mini-apps on Farcaster.

At the moment, the product still feels early, with relatively few apps live.

Minidev also launched its token, $MINI, about a month ago — probably a bit early. That said, @ 0xjenil continues to actively ship and improve the product, so I wouldn’t be surprised to see high-quality apps emerge on Minidev soon.

@ Poofnew— An on-chain Replit.

Poof allows you to create on-chain websites and apps using prompts that are passed to an agent and then distributed across three specialized agents:

UI Agent, Smart Contract Agent, Actions Agent

In July, the team showcased an app built entirely with Poof where two users could duel for the love of a “waifu,” staking SOL in the process.

This kind of project highlights how flexible Poof can be as a playground for experiments.

Games

@ RemixGG_ - a vibecoding platform where you can create games on Base using prompts or play hundreds of games created by others.

Some of them are genuinely solid.

The app is extremely simple: download it on your phone, no wallet required, scroll the feed, play, or build your own game.

This is exactly the type of product that could onboard web2 users who don’t even realize they’re using web3 infrastructure imo.

Dev Tools

@ CodigoPlatform - Codigo is an AI-driven dev platform built specifically for the Solana.

It combines a cloud IDE, purpose-trained AI models, and full build + deploy tooling. Unlike Poof, however, Codigo is hard to classify as a vibecoding platform — you still need to know how to code

Trading Bots

@ pigeon_trade - A trading bot with built-in AI.

Beyond spot, perp, and prediction market trading, you can chat with Pigeon and vibecode custom features, such as:

• Monitoring funding rates on specific DEXs with alerts or automated position opening

• Tracking selected wallets and copy trading

• Searching for arbitrage opportunities on Polymarket

and much more

Final thoughts

We still have very few high-quality web3 vibecoding platforms

Most existing projects are either still in development or haven’t found pmf yet and that’s completely normal.

As mentioned earlier, Virtual Protocol also needed a long time before the narrative fully formed.

Often, products aren’t bad, they’re just too early.

I strongly believe that 2026 will be the year when we start seeing real signals:

Non-devs will build profitable web2 companies through creativity alone, major platforms will integrate vibecoding tools, and user counts and ARR for Cursor, Replit, and Lovable will continue growing exponentially.

As always, degens will be the first.

And a few months or maybe a year later, we’ll look back and compare the beginning of the vibecoding narrative to the rise of pumpfun

The only difference?

Pumpfun gave us hundreds of thousands of new tokens ->

Vibecoding will do the same for web3 apps

Disclaimer:

1. This article is reprinted from [miguelrare]. All copyrights belong to the original author [miguelrare]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.