The fastest way to findAI & Web3opportunities

Three-track hackathon on Telegraph Protocol built on Base. Submit Miners (APIs, models, datasets, or tools), evaluation scripts that rank Miner quality, or applications built on top of live Miners. Part of a 3-hackathon series — Hackathon 1 runs August 17 to September 7, followed by two more rounds through December.

Build concurrent AI agents using Mozaik, JigJoy's framework for running thousands of autonomous agents in parallel instead of sequentially. One open brief, one weekend: build a research swarm, a self-reviewing codebase, a live ops room, or parallel hypothesis testing. Every registered participant gets a discount on Mozaik Cloud and Hyperskill Bootcamps.

Build production-grade AI applications on partner-led tracks from Intel, Qualcomm, SiMa.ai, and Speechmatics, spanning physical AI, on-device AI, edge AI, and voice. Online build is open to everyone; onsite build at the Santa Clara Convention Center during AI Infra Summit is invitation-only and doesn't cover travel or accommodation. Teams of 1-5.

Build a custom Aqua app implementing a sophisticated DeFi position on 1inch's SwapVM stack, with a general track and a separate Continuity track for teams extending existing work. Projects using SwapVM opcodes score higher, and final positions must be demoed through tests or a UI with onchain execution shown.

Build on or integrate any part of the Uniswap stack, including the API, AMM v2/v3/v4, CCA, new v4 hooks, or tooling for the wider ecosystem, with a general track and a separate Continuity track for teams extending existing work. Requires a public repo, a FEEDBACK.md file, and a completed Uniswap Developer Feedback Form submission.

Build on Hedera across four tracks: AI agentic payments with x402, open source contributions to the Hedera Harness, real-world asset tokenization with the Asset Tokenization Studio, or bring an existing Hedera project forward with substantive new work. Each track needs a public GitHub repo and a demo video under 5 minutes.

Build with World's identity stack for verified humans and AI-assisted interactions. Extend an existing project with AgentKit to distinguish real human-backed agents from bots, open only to Continuity Track participants, or build a Selfie Check flow for risk, eligibility, or abuse prevention. Both tracks require a working app tested in the World ID Sandbox plus a written feedback document on the developer experience.

Build privacy-enhanced applications using Midnight's zero-knowledge technology. First wave of a three-wave grant program, part of a 12,500 USDT total pool. Submit before the wave deadline to be judged and eligible for the grant.

Details limited — description of what to build was not available at time of listing. Related to Commonsmade's VibeFi. Includes a "Vault" prize structure paying out $20,000 plus 80% of all revenue generated by the winning entry, in addition to standard placed prizes.

Three-week hackathon to build and ship any kind of AI agent — trading, social media, research, content, community tools, or anything new. Seven winners share the prize pool. Every entry receives an AI vetting score at submission plus community upvotes from registered builders. Judged on real work and working demos.

Build a Mini App using the Nimiq Pay Mini Apps Framework, which handles wallets and payments so you can focus on the product. Open to developers, AI builders, vibe coders, and indie hackers, solo or in teams. Apps must be open source and can span games, productivity tools, marketplaces, or social experiences.

Build the best integration of KeeperHub, a deterministic execution layer for onchain agents, into a live project with real users. A separate bounty rewards shipping an open source feature to the KeeperHub repo itself, which stacks with the main track. Submissions need a source code link, a demo video, and a link to a transaction executed through KeeperHub.

Build an AI agent on Solana, trader, tool, or product, then tokenize it on ClawPump or EasyA Kickstart by Sept 19 as your entry ticket. Judged live over six weeks on streams by a panel including Ansem, Solana Foundation, Delphi, and Haun. Most of the prize pays out in vesting $ANSEM tokens rather than upfront cash, and entry requires deploying a real token on-chain.

Build a live, API-backed AI or crypto capability and prove it works through technical, reproducibility, and security review, with a path to becoming a standardized MCP tool on OKX.AI for ongoing paid-call revenue. Choose the General Challenge for any useful capability, or the OlaXBT x X-Agent Trading Challenge using the OlaXBT Nexus MCP for strategy development and backtesting. On-chain security and auditing projects are not eligible. You don't need to build an MCP server yourself before submitting.

Build on emma, a multi-cloud GPU infrastructure platform spanning 11 providers and 23 GPU models, controllable via prompt or API. Choose Builders (2-3 weeks, teams of 1-5, ship a working product using emma's API/SDK) or Showcases (a weekend, solo-friendly, a prompt-to-result demo under 2 minutes). Applications are curated, show a shipped infra repo or build thesis to get in. Free GPU credits for the full hackathon.

Three-month challenge to build a real-world application on XRPL Mainnet, attract users, and generate on-chain volume. Weekly leaderboard tracks progress. Prize categories reward most users, highest volume, hitting 300 active users, and a jury grand prize for best overall project. Top projects can enter the Aquarium incubator.

Build a new full-stack app with Convex as the backend, Firecrawl for data, and AgentMail for inbox capability, using Codex or any agent with the Convex plugin. Every participant gets 20k in Firecrawl credits during the build. Judges want real apps a person would use, not developer tools.

Build an AI-powered solution to a real-world problem, open to any domain: generative AI, computer vision, NLP, healthcare, sustainability, productivity, or developer tools. Judged mainly on innovation and creativity of the idea.

Pitch a voice AI concept for Banking & Insurance or Government Services in Dubai, then build a working agent on the ElevenLabs platform. Stage 1 is a strict five-page Idea Canvas due Sept 23, no other formats accepted. Top 20-30 teams get platform access to build Sept 30-Oct 14, and 5 finalists pitch live at Demo Day in Dubai on Oct 26/27. Must be multilingual (English and Arabic minimum) and compliant with ElevenLabs' terms.

Build an AI project alongside WeAreDevelopers World Congress North America in San José. Take part fully online, or apply for on-site days for additional prizes and a chance to showcase live, travel and accommodation not covered. All online participants get a ticket to the World Congress. Challenge tracks, tech stack, and prize pool were still to be announced as of listing.

Build on X Layer for tokenized stocks and real-world assets, or on OKX AI for agent services and AI-native businesses. Apply by Sept 11 as a team of up to 4 or solo, then build online Sept 17-25 with submission due Sept 25. Selected teams present live at the finale in Singapore on Oct 6, though remote teams can also win. OKX does not cover travel for finalists.

Build with IBM Bob 2.0, an AI development partner that works with full repository context to help developers reason through logic and navigate complex codebases. A 48-hour online build, solo or in teams, judged on application of the technology, presentation, business value, and originality. Specific challenge tracks and submission requirements were still to be announced as of listing.

A machine learning competition on agricultural data, not a product-building hackathon. Registration closes Sept 23, submissions due Sept 30.

Build a voice AI agent on AssemblyAI, either through the end-to-end Voice Agent API or the Realtime Speech-to-Text API with your own orchestration and LLM. Month-long build window, registration stays open the whole time. Judged on application of technology, presentation, business value, and originality.

Global mobile hackathon for builders launching real apps to the App Store or Google Play. Compete across 21 prize categories including Grand Prize, Best Game, Design Award, Build in Public, and Next Gen (students). Grand Prize winner gets cash, a Times Square billboard feature, and an invitation to App Growth Annual in NYC. Must use RevenueCat for in-app purchases and subscriptions.

A research contest on Kaggle to advance general AI reasoning against the ARC-AGI-3 benchmark. No internet access during evaluation, and GPT or Claude APIs are excluded. Cash prizes require open-sourcing your solution by the deadline.

Deploy a paid x402 endpoint on Algorand Mainnet — charge per data request, compute job, action, or verification — and generate real on-chain usage. Top 50 on the leaderboard compete for 10 finalist spots who pitch live at Devcon 8 India. Judged on real usage, use case quality, technical execution, and long-term potential.

Build with Arbitrum, the Layer 2 ecosystem, online. Existing products are allowed as long as meaningful new work happens during the buildathon. Part of the Open House series alongside the Dubai edition, with grants and a promising-products prize sitting alongside the placed cash awards.

Five-week business-first online build program to learn Canton, form ideas, and ship production-style MVPs. Single track: Real-World Asset and Business Workflows — build end-to-end workflows for issuance, state changes, transfers, and audit. Top 5–10 finalists compete in a live Grand Final on October 14. Strong post-hackathon ecosystem support for fundraising and pilot launch.

Build an AI agent that does real work for a real user, industry, or community, not a demo. No fixed theme, choose from 8 tracks spanning code-shipping agents, ops copilots, research agents, public sector, commerce, multimodal field agents, and safety/red-teaming. 72-hour online build, teams of 1-5, with mentor access and live workshops throughout.

An online hackathon bringing together builders and founders across blockchain ecosystems to launch breakout startups. Prize tracks will focus on select chains, each backed by an ecosystem partner, plus a general prize pool for the best products across all chains. Full details on judges, prize tracks, sponsors, and dev resources release on Sept 14, the same day the hackathon begins.

A six-week global online hackathon to build the next generation of onchain startups on Monad. Four tracks: Onchain Finance & Trading, Consumer Products & Payments, Social/Attention/Culture, and Trust, Identity & AI Infrastructure, each paying $30,000 split evenly between 3 teams, plus a $25,000 grand champion across all tracks. Extensive sponsor bounties on top from Kuru, Agora, Privy, Nansen, Chainlink, and others. Open to solo builders and teams, no prior onchain experience needed, submitted work must be built during the six weeks.

Build privacy-enhanced applications using Midnight's zero-knowledge technology. Second wave of a three-wave grant program, part of a 12,500 USDT total pool. Submit before the wave deadline to be judged and eligible for the grant.

Build AI agents and high-performance AI applications entirely in the cloud on AMD GPUs and the ROCm software stack, no local hardware needed. Requires signing up for the AMD AI Developer Program to be approved. Online build for everyone, with an on-site phase across three Italian cities, Rome, Milan, and Imperia, for approved participants, travel and accommodation not covered.

Design and build a new subnet for the Bittensor network, from mechanism design to a working testnet implementation. Define what work miners perform, how validators score it, and how the incentive system rewards contribution. Submit progressively: a subnet proposal by Sep 20, then a full testnet implementation, GitHub repo, demo video, and final pitch by Oct 19. Strong projects may get incubation support through Bitstarter or interviews with Bittensor accelerators.

Formerly the LLM Hackathon for Materials Science and Chemistry, now expanded to the physical sciences and mathematics. Build LLMs, autonomous agents, datasets, benchmarks, models, or scientific software against real research problems, in teams of 3 to 10, with mentors from academia and industry. No cash prize, judged on impact potential, innovation, scalability, and domain relevance.

Real genomic and clinical data from a child living with Mosaic Variegated Aneuploidy, an ultra-rare genetic condition, shared by the family to find answers. Track 1: predict the causal genetic variant from FASTQ and VCF data, scored automatically on a live leaderboard. Track 2: propose a repurposed existing medication based on the disrupted biological mechanism, judged by an expert panel. Data access is controlled and requires an approved request; all outputs are released under CC BY 4.0.

Build using Nebius AI cloud infrastructure with at least one NVIDIA open-source model. Includes city-level and feedback awards alongside the main prizes; a Jetson Orin Nano is offered as hardware, not cash.

Build a fully deployed, revenue-generating SaaS platform with a live monetization engine, not a prototype. Open to all ages and occupations globally, despite Devpost's student-only platform tag. Submissions need a live public deployment, a public repo, a working MVP, and a demo video, judged on architecture, compliance, market fit, monetization, UI/UX, and pitch.

Build browser-native AI tools and 3D experiences that transform natural language into interactive worlds. Projects must run in the browser and turn plain language into 3D scenes and working logic. Covers AI-generated 3D scenes, intelligent NPCs, procedural worlds, and real-time AI interactions using any web-based 3D stack.

Build privacy-enhanced applications using Midnight's zero-knowledge technology. Third and final wave of a three-wave grant program, part of a 12,500 USDT total pool. Submit before the wave deadline to be judged and eligible for the grant.

Four-month hackathon to build and launch live products on QIE Mainnet — not demos. Five tracks: DeFi, AI and Autonomous Agents, Consumer Apps, Commerce and Real World, and Developer Infrastructure. Prizes split 50% on win and 50% released after hitting 90-day traction milestones covering wallet growth, transaction volume, and revenue.

A 3-month AMD-powered developer growth program, individual participation only, running monthly build challenges from Sept 1 to Dec 1. Earn XP through building, learning, community support, and referrals to climb from Bronze to Legend rank. Cash prizes go to the first 3 builders to reach Legend, which requires 10,000+ XP plus a final challenge unlocked near the end of the program.

Build with Arbitrum, the Layer 2 ecosystem, across DeFi, gaming, social, tools, or anything new. Use Stylus or Solidity, deploy on Arbitrum One or your own Orbit chain, and bring an existing product or start fresh. Top projects get exposure to the Arbitrum ecosystem and connect with regional leaders in Dubai. Starts Oct 31.