From August 15–18, developers from around the world gathered in New York City for a landmark hackathon track at ETHGlobal New York, hosted at the iconic Metropolitan Pavilion.
With $10,000 in prize money and early access to state-of-the-art tools, the hackathon was a small but incredibly promising glimpse into the future of decentralized artificial intelligence.
The track was presented by the Artificial Superintelligence (ASI) Alliance, a collaboration between the three alliance members – SingularityNET, Fetch.ai, and Ocean Protocol – powered on-site by the Fetch.ai Innovation Lab.
The focus? To challenge developers to create modular AI agents that do more than execute code. These agents perceive, reason, act, and collaborate across decentralized systems.
Participants were invited to use Fetch.ai’s uAgents framework, deploy to Agentverse (the ASI-wide registry for agent orchestration), enrich agents with MeTTa symbolic reasoning and structured knowledge, and pull in decentralized data using Ocean Protocol. On top of that, teams had access to ASI:1, a conversational AI interface that let users interact with agents through natural language.
The energy inside the Pavilion was electric! Some participants arrived with deep experience in Web3 and AI. Others were fresh to decentralized agents but brought bold new perspectives. For three days, teams debated, sketched, coded, broke things, and rebuilt them, fueled by the shared belief that the future of AI should be open, interoperable, and beneficial to all.
As Sana Wajid, SVP at Fetch.ai Innovation Lab, put it:
“Hackathons like ETHGlobal New York really give us a glimpse into the future – the one that is shaped by creative developers building meaningful AI applications. We’re grateful to have the opportunity to present this event with the Artificial Superintelligence Alliance and see the ideas come to life. From autonomous agents to discoverable multi-agent systems, the projects showcased the real-world potential of interoperable intelligence through Agentverse and ASI:One. These projects are preview of the agent-driven ecosystem of tomorrow, and we’re just getting started.”
Ibby Benali, Ecosystem Marketing Lead at SingularityNET:
“We’re incredibly proud of how the ETHGlobal NYC hackathon unfolded—thanks in no small part to the stellar execution by the Fetch.ai Innovation Lab. Their commitment to developer empowerment and decentralized innovation was evident at every step. This collaboration truly showcased the strength of the ASI Alliance in action, and we’re excited to build on this momentum together.”
Mihai Scarlat, Product Lead at Ocean Protocol:
“For us, ETHGlobal New York was all about building and putting our stack into the hands of developers, stress-testing it in real use cases, and hearing their feedback first-hand. That’s how we grow stronger as an ecosystem. We’re grateful to have stood alongside our ASI Alliance partners, and a special thank-you to Fetch Innovation Lab for driving the developer relations that made this possible.”
Kshipra Dhame, Developer Advocate at Fetch.ai:
“At ETHGlobal NYC, what struck me most was how quickly developers embraced the full ASI Alliance stack. Developers used Fetch.ai’s uAgents to build AI agents, deploy them on Agentverse, and make them accessible through ASI:One. Some extended their agents with symbolic reasoning using Singularity.Net’s MeTTa, while others published live datasets to OCEAN for reuse and composability. The winning projects showed that these technologies aren’t isolated tools — they already fit together as building blocks of a real agent-driven ecosystem.”
The result of the hackathon was an impressive showcase of projects that didn’t just meet the challenge but redefined what’s possible when symbolic reasoning meets autonomous agents in a decentralized environment.
Today, we’re excited to showcase the standout projects that impressed the judges and inspired everyone watching.
🥇 1st Prize – Etherius
Best use of Artificial Superintelligence Alliance
Built by: Mikhail Vaysman
What it does:
Etherius built an AI-powered NFT intelligence agent that connects natural language queries to real-time NFT market data via OpenSea.
Built with Fetch.ai’s uAgents, it allows users to ask questions like “What are the top trending gaming NFTs today?” and get structured, human-readable answers; complete with floor prices, trading volume, and OpenSea links.
Why it stood out:
Etherius nailed the user experience aspect. It bridges NLP with live market data and presents it in a way that’s instantly usable. The payment integration using the x402 Fetch.ai agent was smooth, and the “buy NFTs on the go” functionality demonstrated how conversational agents can power real commerce.
Explore their GitHub Repo.
🥈 2nd Prize – MintCondition
Best use of Artificial Superintelligence Alliance
Built by: Alex Lu, Lindsay Zhang
What it does:
MintCondition is a multi-agent NFT appraisal system that uses multiple LLMs to independently evaluate NFT prices. Then it uses MeTTa symbolic reasoning to reach consensus, with support from the ASI:One system to structure and validate outputs. You can ask it to appraise a specific NFT, watch the agents reason in real time, and get a confidence-scored price range based on agent agreement.
Why it stood out:
This is one of the most technically ambitious projects we’ve seen. The use of streaming agent logs, real-time consensus reasoning, and WebSocket infrastructure gave a transparent view into what each agent is thinking. It felt like watching a boardroom of AIs negotiate a fair price—and that’s exactly the kind of system we envision for decentralized intelligence.
Explore the GitHub Repo.
🥉 3rd Prize – Waterfall
Best use of Artificial Superintelligence Alliance
Built by: Divyansh Goel, Dayitva Goel
What it does:
Waterfall connects Fetch.ai agents to the OpenSea MCP server to fetch real-time NFT and token data. Then it publishes structured datasets through Ocean Protocol, making this data accessible for others to buy, use, or integrate. It also anchors the data to a smart contract, adding transparency and traceability to what’s collected.
Why it stood out:
Waterfall took on the ambitious goal of turning real-time agent data into reusable, decentralized datasets. By combining agent infrastructure, Ocean Protocol publishing, and on-chain verification, it showed a full-stack vision of how AI-collected data can become a shared, monetizable resource.
Explore the GitHub Repo.
Honorable Mentions
Synapse
Synapse is a DeFi portfolio manager built with Streamlit, combining wallet tracking, performance analytics, and token swapping via Uniswap and CCTP. It features an AI assistant powered by LangGraph and GPT-4o-mini that delivers portfolio insights, yield strategies, and risk assessments. Clean UI, smart features, and a surprisingly complete tool for a weekend build. The groundwork was there for a powerful coordination framework, and we’d love to see it evolve further.
Team: Tirth Joshi, Harsh Bajaj
LoveFi
LoveFi took a creative spin on decentralized AI, blending dating, staking, and agentic interaction. Users could create relationships, while friends placed bets on whether the couple would stay together. Behind the scenes, LoveFi used uAgents with a local language model, integrated ASI reasoning, and a custom chat interface to bring a playful, emotionally aware experience to life. It was fun, bold, and hinted at how social DeAI applications might evolve.
Team: Daniel Wang, Faye Hall, Shanthan Sudhini, Garrgi Pathak
The Takeaways (and What’s Next!)
Across all submissions it was clear to our board that developers are starting to treat AI agents not just as tools, but as collaborators.
The best projects didn’t just use GPT or fetch data, they composed multiple intelligent systems, Fetch.ai’s uAgent framework, structured reasoning protocols like MeTTa, and decentralized coordination layers.
This hackathon showed us what’s possible when you combine symbolic reasoning with neural models, plug them into real market protocols, and build with openness in mind.
We’re proud of every team that showed up and built. Whether you took home a prize or not, the work you did is part of something much bigger.
Stay tuned for the next one. We’re just getting started.
