About this Event
The era of simple chatbots is over. Real AI applications need to access live data, write their own code, autonomously manage their own state, and run reliably in the cloud.
The Agent Forge AI Hackathon gives you the ultimate stack to ship production-ready AI agents in one day. Assemble a team of one to six people — or go solo — and build, test, and deploy a working Agent Forge solution together. We are moving beyond local prototypes. If it doesn't deploy and manage real data, it doesn't count.
Team Format
Form a team of 1–6 people. Teams can blend different skill sets — developers, designers, business minds — to tackle bigger, more ambitious projects. Here are some directions your team could take:
- A custom UI to manage multiple Agent Forge instances — build a unified dashboard for orchestrating and monitoring several agents at once.
- Custom tools or skills for a single Agent Forge agent — extend what one agent can do by shipping new capabilities it can call on.
- A collaborative, cross-role project powered by Agent Forge — combine business development, design, and engineering: one teammate builds a pitch deck, another designs the UI, and a third wires it all together through Agent Forge.
You're free to come up with your own idea. The only requirement: your project must be built on Agent Forge and deployed live by demo time.
The Challenge
Build an autonomous Agent Forge assistant or data-driven application that solves a real-world problem.
Unlike standard hackathons where you build UI shells, this event focuses on functionality, live data, and true autonomy. Your Agent Forge agent should be able to gather intelligence (Scraping), autonomously process information, build with professional-grade tooling (Coding), and live on the web (Deployment).
The Stack (Your Arsenal)
You will have access to premium credits and tools from our innovation partners to build your Agent Forge solution:
: The open-source AI backend. Define an agent, add one decorator, and you have a production-ready microservice with async execution, shared memory, and live observability - fully deployed. Multi-agent coordination is one line of code. No extra infrastructure, no glue code. Your agents reason. Your infrastructure governs
: The browser action engine for AI agents. Equip your agents with up-to-date action manuals and verified DOM selectors to operate any website instantly. Eliminate guesswork and brittle scraping, empowering your agents to seamlessly handle modern web apps, SPAs, and complex workflows with unbreakable resilience and speed.
: Give your assistant the ability to browse the web. Build custom web-scraping "skills" into your agents to fetch real-time data, monitor trends, or aggregate research automatically using reliable, large-scale infrastructure.
: The official platform for the Qwen model family. Access state-of-the-art open-source and proprietary models optimized for deep reasoning and agentic coding. Empower your agents to process multimodal information and seamlessly execute complex real-world workflows.
: The agentic coding platform built for real-world software development. Move from simple assistance to true autonomy using features like Expert Panel Mode to assemble a specialized AI engineering team, and Repo Wiki to instantly surface implicit codebase knowledge.
: Spin up a dedicated server on Zeabur to deploy your app, debug DevOps questions with Zeabur agent and use the web interfaces with a single click, ensuring your project is live, accessible, and scalable.
: Creators of the frontier GLM model family. GLM‑5.1 is its latest frontier large language model, built as a SOTA open‑source model for long‑horizon tasks with Opus‑level reasoning depth. Build websites, create slides, analyze data, and get instant answers.
The Workshops
We will host hands-on sessions to get you up to speed with the tech stack.
Prizes
All winning teams will earn a share of exclusive hackathon prizes, including sponsored credits and tools from our partners to keep building and scaling their projects.
Agenda
- 10:00 AM — Check-in & Networking
- 10:30 AM — Hackathon Introduction
- 10:45 AM — Product Workshop
- 11:30 AM — Team Formation & Project Setup
- 12:30 PM — Lunch
- 4:30 PM — Submission Deadline / Live Demos
- 5:30 PM — Winners Announced + Networking
- 6:00 PM — Close
Who Should Come?
You want to build AI agents that access the real internet, not just static datasets. You want to deploy something live, not just run it on localhost. You want to connect with the Silicon Valley AI community and the AI Builders network.
Whether you're a developer, a designer, or a business builder — if you want to ship a real AI project in a day with a team, this is your event.
Organized By
: A premier community of AI builders, developers, and innovators in Singapore, Tokyo, and San Francisco Bay Area. Prototype real products with fellow builders, ship features weekly, and trade production-tested playbooks.
Beta Fund is an early-stage venture fund based in Palo Alto, backing the builders shaping the AI era. We partner with exceptional technical founders at the inception stage. Our focus spans agentic AI, AI-native systems, and AI-powered vertical SaaS—the foundational layers of the next software epoch.
: Llama Ventures is an early-stage venture firm backing ambitious founders building AI-native companies by providing deep technical expertise and strategic capital across the intelligence stack to transform innovation into market reality.
Limited spots available. Register now.
Event Venue & Nearby Stays
Llama Ventures, 880 West Maude Avenue, Sunnyvale, United States
SGD 0.00
