Full-Stack Blockchain Developer | Solidity | Next.js | React | TypeScript | Web3 | Building the future of decentralized applications

Joined December 2024
270 Photos and videos
Pinned Tweet
In 5 years, I didn't just build in Web3. I moved the ecosystem. Mass adoption. Real-world problems solved. My fingerprints on how this space grew. Oh — and the body? Built different too. Not a dream. A deadline.
3
5
50
1,409
ARIA is turning out crazier than I expected. Next up is @MetaMaskDev Smart Wallet Kit integration and @AskVenice as the main brain behind ARIA
Anticipate!!!
2
4
69
Anticipate!!!
I'm going to be pausing my current project for a bit because a hackathon appeared out of nowhere 😭 But honestly, I'm excited because I'm building one of the coolest things I've worked on yet. Introducing ARIA. ARIA is an open marketplace and discovery layer for AI agents. The idea is simple: A user comes to ARIA and describes what they need in plain English. Instead of relying on a single AI model, ARIA understands the request and searches a global registry of specialized AI agents built and published by developers from anywhere in the world. It then routes the task to the most suitable agent (or combination of agents) for the job. Developers can build virtually any type of agent: • Creative agents for image, video, audio, and content generation • Research agents with access to proprietary datasets • Healthcare agents connected to hospital databases • Legal agents connected to law firm knowledge bases • Financial agents with access to specialized market intelligence • Or entirely new categories that don't exist today The best part is that developers don't have to build their agents on ARIA. They can host them anywhere, use any model they want, connect any APIs they want, and use whatever infrastructure they prefer. All they need to do is register their agent on ARIA and integrate a MetaMask-powered payment facilitator so their agent can receive payments whenever it's used. ARIA's job is discovery, routing, payments, and coordination. For privacy, ARIA uses @AskVenice as the platform intelligence layer. When a request comes in, ARIA can analyze it, coordinate with specialized agents, aggregate their responses, and generate the final output using @AskVenice. This means users who care about privacy don't have to worry about their prompts, requests, or data being stored and reused by the platform AI. If ARIA can't find a suitable agent for a request, the platform AI handles it as a fallback. That request is then logged as unmet demand, creating a real-time signal showing developers exactly which types of agents users are actively looking for but don't yet exist. This turns user demand into a roadmap for builders. Using @MetaMaskDev Smart Wallet Kit, agents can be given controlled access to funds, allowing developers to charge per execution while users define spending limits before a task begins. The long-term vision is simple: - Anyone can build an agent. - Anyone can publish an agent. - Anyone can monetize an agent. - Anyone can discover and use the best agent for a task. Not another AI assistant. A marketplace for intelligence. I'll be sharing daily build updates as I bring ARIA to life
1
1
6
127
I'm going to be pausing my current project for a bit because a hackathon appeared out of nowhere 😭 But honestly, I'm excited because I'm building one of the coolest things I've worked on yet. Introducing ARIA. ARIA is an open marketplace and discovery layer for AI agents. The idea is simple: A user comes to ARIA and describes what they need in plain English. Instead of relying on a single AI model, ARIA understands the request and searches a global registry of specialized AI agents built and published by developers from anywhere in the world. It then routes the task to the most suitable agent (or combination of agents) for the job. Developers can build virtually any type of agent: • Creative agents for image, video, audio, and content generation • Research agents with access to proprietary datasets • Healthcare agents connected to hospital databases • Legal agents connected to law firm knowledge bases • Financial agents with access to specialized market intelligence • Or entirely new categories that don't exist today The best part is that developers don't have to build their agents on ARIA. They can host them anywhere, use any model they want, connect any APIs they want, and use whatever infrastructure they prefer. All they need to do is register their agent on ARIA and integrate a MetaMask-powered payment facilitator so their agent can receive payments whenever it's used. ARIA's job is discovery, routing, payments, and coordination. For privacy, ARIA uses @AskVenice as the platform intelligence layer. When a request comes in, ARIA can analyze it, coordinate with specialized agents, aggregate their responses, and generate the final output using @AskVenice. This means users who care about privacy don't have to worry about their prompts, requests, or data being stored and reused by the platform AI. If ARIA can't find a suitable agent for a request, the platform AI handles it as a fallback. That request is then logged as unmet demand, creating a real-time signal showing developers exactly which types of agents users are actively looking for but don't yet exist. This turns user demand into a roadmap for builders. Using @MetaMaskDev Smart Wallet Kit, agents can be given controlled access to funds, allowing developers to charge per execution while users define spending limits before a task begins. The long-term vision is simple: - Anyone can build an agent. - Anyone can publish an agent. - Anyone can monetize an agent. - Anyone can discover and use the best agent for a task. Not another AI assistant. A marketplace for intelligence. I'll be sharing daily build updates as I bring ARIA to life
Building an on-chain crowdfunding app, but not the regular kind. Builders define milestones upfront, and contributors vote to release funds at each stage. No disappearing with the treasury, no blind trust, just transparent execution and accountability. Lets see how the build goes
1
6
211
Cracked Dev retweeted
I need help guys!! What makes an AI Agent truly an AI Agent?
2
3
61
Building an on-chain crowdfunding app, but not the regular kind. Builders define milestones upfront, and contributors vote to release funds at each stage. No disappearing with the treasury, no blind trust, just transparent execution and accountability. Lets see how the build goes
1
9
201
Had an amazing time building during the hackathon, so sad that I could not be in Milan physically for the final day.
Replying to @lablabai
Prizes from @Vultr : 1st - Deals Machine: The Autonomous Sales Agent by View 1 Studio 2nd - ContextGuard Live Decision Helper by Revived 3rd - Cascade AI by Cracked Dev
1
3
139
Cracked Dev retweeted
Prizes from @Vultr : 1st - Deals Machine: The Autonomous Sales Agent by View 1 Studio 2nd - ContextGuard Live Decision Helper by Revived 3rd - Cascade AI by Cracked Dev
1
2
4
362
What the hell is this error @greyfinance??????
5
4
343
Won a hackathon and just got the “send your payment details” email 😭 They said PayPal or wire transfer… meanwhile we know how PayPal in Nigeria is, and this is my first time dealing with wire transfers. Tried Grey but faced issues, now looking at @clevabanking. Thoughts?
8
9
564
Had no idea hackathons took this long to pay 😭
1
7
181
Definitely my best month so far 😭🔥
7 days ago I had never built an AI agent in my life. Today I shipped Cascade AI — a multi-agent B2B onboarding platform powered by Gemini 2.5 Flash. Here's the full breakdown🧵👇
5
1
22
492
Cracked Dev retweeted
Need help from university students 🙏 If you’re currently in university or recently graduated, please help me fill this short survey. Would really appreciate reposts and shares to student groups/channels too 🙏 forms.gle/N5YJtF6bA1cSe8EP9 #NaijaStudents #StudentHustle #BUK #Unigram
1
4
71
7 days ago I had never built an AI agent in my life. Today I shipped Cascade AI — a multi-agent B2B onboarding platform powered by Gemini 2.5 Flash. Here's the full breakdown🧵👇
3
2
13
684
First AI agents. 7 days. Biggest lesson: the hard part isn't the code. It's understanding when to let Gemini decide vs when to guide it with structure. Competing in @lablabai AI Agent Olympics.
1
1
56
Live site: cascadeai.live/ GitHub: github.com/YusufsDesigns/Cas… Project page (please like/support): lablab.ai/ai-hackathons/mila…

1
1
56