Over the past few months, we’ve run 9 in-person Mini-Hackathons across different cities—from Shenzhen to Kuala Lumpur. At each event, developers formed teams, built projects, created demos, and presented their work on stage. Now, we’re bringing that momentum online for a one-month grand finale Hackathon.

The idea is simple: build a working, demo-ready product prototype using AI. It doesn’t need to be perfect—but it does need to exist and be something you can actually show.

We’re organizing the event around 3 Sponsor Challenges. Each partner will provide a themed problem statement. You can pick the one you’re most interested in and build using their APIs, SDKs, and technical stack. You might create an AI agent, a developer tool, a data product, or any kind of AI-powered application—the direction is wide open.

There will be workshops and office hours along the way. You can join as a team or go solo. One month later, we’ll host a Demo Day where selected projects will be presented live, with prizes and awards (including MacBook, DJI Pocket 4, and cash prizes for each track).

This is not about listening to talks—it’s about shipping real things. 😄

📅 Time & Format

Fully online

🗓️ Timeline

Development Phase (June 22 – July 9)

• 30–45 minute workshops / office hours hosted by each sponsor to help you get started with their tech stack

• Form teams freely, or work solo if you prefer

• Daily community check-ins and progress updates to keep momentum going 😏

Submission Period (July 9 – July 16)

• Submit your demo video, GitHub repository, and technical documentation

• Don’t wait until the last minute—you know how it goes

Demo Day (July 17 / July 19)

• Selected projects will be presented on stage

• Each presentation: 5–8 minutes

• Q&A with judges and live discussion

• This is the highlight of the entire event 🌟

Results Announcement (July 27 – July 28)

• Overall winners + Sponsor Challenge Awards announced

• Prize distribution and potential follow-up collaboration opportunities

🔛 Three Tracks

Track 1: Make It Agent-Payable

Sponsor: KITE AI

AI agents today have strong execution capabilities, but lack trusted autonomous action and payment mechanisms. _Kite Agent Passport_provides any agent with verifiable identity, owner-authorized spending limits, and stablecoin payment support, with all operations fully traceable on-chain.

This track is fully open-ended: you may build any project that enables agent autonomous action + self-payment based on Kite Agent Passport, covering demand-side, supply-side, on-chain scenarios, and consumer-facing applications.

Reference directions:

• Agents that autonomously pay for APIs/data/compute resources within preset budgets

• Upgrade existing API/MCP servers or apps to support agent identity verification + pay-per-use billing

• Inter-agent hiring and paid collaboration systems

• Agents executing on-chain operations within owner-granted spending limits

• Consumer AI assistants with controlled wallets supporting subscriptions/shopping/bookings

• Agent payment quota management, bill checking, and audit tooling

Track 2: Build with 0G Private Computer

Sponsor: 0G Labs

Build verifiable inference applications using decentralized AI compute. 0G Private Computer is a decentralized compute network for AI agents, offering verifiable inference services with on-chain attestation and traceability for all model calls. The platform provides access to 15 mainstream large models at up to 80% lower cost than official pricing.

Track goal:

Call AI models via the 0G Compute API to build a runnable Web3+AI product demo, exploring integration scenarios for AI and blockchain.

Basic requirements:

• Integrate at least one model from 0G Private Computer via the official API

• Submit a runnable demo (Web App/CLI/Bot/Agent)

• Provide a public GitHub repository with a clear README

• Share a 3-minute product demo video

  • Bonus points for: Highlighting verifiable inference value, integrating Web3 scenarios (DeFi/GameFi/DAO/NFT/on-chain agents), using the self-developed 0GM-1.0-35B-A3Bmodel, or enabling multi-model collaboration.

  • Not accepted: Solutions that only call third-party AI APIs without 0G integration, non-runnable code, or pure whitepaper concepts.

Track 3: Gonka: AI for Society

Sponsor: Gonka

Leverage Gonka’s decentralized inference network to build inclusive, trustworthy, neutral public AI tools that serve real-world social needs.

Against the backdrop of widespread AI-generated misinformation, centralized platform bias, and monopolistic limitations of public AI tools, this track focuses on real-world public value delivery. We encourage participants to build accessible, open-source public AI tools for ordinary users, using decentralized AI to solve practical social problems.

Priority is given to Decentralized Truth Engines: Rely on multi-model inference to verify the authenticity of news, social posts, and digital assets in real time; leverage Gonka’s decentralization to avoid centralized bias, and provide neutral, transparent, on-chain attested information verification solutions to counter deepfakes and misinformation.

Mandatory technical requirements:

• All AI inference and verification logic must run through the official Gonka Router gateway (gonkarouter.io)

• Multi-model cross-consensus is encouraged: integrate at least two different LLMs (e.g., Qwen, Kimi, DeepSeek) for cross-validation to ensure neutrality.

  • Core required features:

• Support input of URLs, tweets, text snippets, or images for verification

• Decentralized verification combining real-time web data and model knowledge bases for multi-dimensional authenticity analysis

• Output a standardized 0–100% _Truth Score_with fully traceable reasoning explanations

• Transparent UI showing verification results and corresponding Gonka Request IDs for auditability

  • Support for participants: Unlimited free Token quotas are provided for development and testing during the hackathon.

🙋 Who Should Join?

• Developers / builders who want to turn ideas into working, demoable prototypes—not just concepts

• People interested in specific Sponsor Challenges who want to build with their APIs and possibly win prizes

• Anyone looking to find teammates and build together 🤝

• Builders interested in AI agents, developer tools, data products, AI + Web3, or consumer AI applications—the scope is intentionally open

• Anyone who enjoys producing something tangible within a month

There are no strict barriers. If you want to build something, you’re welcome to join 😊