The Agent Arena Hackathon

Many websites you use every day, like Luma or Facebook Marketplace, have no API, so your AI agent has no way to use your account. Today the only option is a…

AI models on CPUs are held back by generic kernels. Nusa is an agent that writes faster ones for your exact hardware, and it's safe to leave running on its…

Only about 2% of information is digitized. The vast majority of information by humans exists in undigitized manuals, CAD diagrams, unarchived research specs…

This will take some untrusted github repo, and a benchmark python script displaying a performance regression, and it will make your code fastr! It makes…

Glasshand lets AI agents running on Vultr operate real iPhones by reading their screens and using human-like taps. Unlike virtualized devices, these phones…

Cerberus: evidence backed repo verification on throwaway Vultr VX1s. You hand it a repo. It clones it read-only, then boots it on a disposable VX1 in your priv…

Autonomous agents are being handed wallets, and every tool call that fetches external content is a prompt-injection surface — the most damaging thing an…

for Challenge 2 🤖 Powering the Future of Work with Scalable, Deployed AI Agents. Problem is businesses they don't know how to create good ai agents and…

Pengwin puts event coordination and office operations in one Slack channel. Mike asked Concierge to plan a Salesforce Park meetup for 30 founders, source 24…

GhostRange is an autonomous cyber-defense proving ground built for safe agent execution. Instead of allowing an AI agent to experiment directly on production…

AI agents can write analysis code, but running that code introduces two challenges: incorrect results and uncontrolled execution. Market Watch Commander…

ChatGPT learned by reading the internet. Robots can't: every robot skill today is recorded by hand, one task at a time, and the new robot models (VLAs) are…

Small businesses win government work by answering RFPs, and most never try: finding the right solicitation, reading a fifty-page packet, and writing a…

Upgrade Chamber is a live demonstration of one Python dependency upgrade, tested against a repository's own test suite, repaired by a bounded agent when the…

A web-based agent that does real work (writes and runs code) where every action happens inside a throwaway, hardened sandbox on a Vultr VM

GeoScope helps planners and community teams explore where a clinic, library, school or community centre could fit and whether it would improve access. Users…

Sim2Sale allows you to run thousands of simulated agents around product offerings. Understanding buying intents, product variations, and optimizing sales…

MiniAVO is an evolutionary search over Triton GPU kernels. Each generation an LLM (OpenRouter or Vultr Serverless Inference) rewrites the current best kernel…

Judges Password: pas89-auar7-gju9z-msyn9 Ontofill is an open-source engine. You give it a one-sentence question, and it runs five phases on its own: it writes…

ShiftOps is a facility operations platform built for Vultr’s Agent Arena Challenge. It runs on a Vultr virtual machine, which powers the public web app…

I've merged more than 140 fixes into ML libraries like DeepSpeed, Unsloth and sentence-transformers, and about two-thirds were silent math bugs: no crash, CI…

Product teams need to understand how people use their personal AI assistants (e.g. ChatGPT, Muse, Grok Bot) and where they struggle, without breaching user…

Mira Picks is an autonomous AI fashion publication built around a multi-agent editorial desk and an agent-driven production pipeline. Miranda, the…

Kettle runs the sales-to-cash and procure-to-pay cycle for a mid-size IT hardware reseller using three coordinated AI agents and an AI orchestrator over a…

After a flood, families have to prove every loss to get aid, and in the same week scammers target them with fake relief websites. Professionals document…

X/Mine runs self-driving haul trucks in an open-pit mine on Vultr, with a crew of AI agents on Vultr Serverless Inference handling driving, traffic and…

"Blast Radius Zero" - Self evolving multi-layer harness for secure and productive agentic execution

Tieout is AI month-end close for accounting firms: one click reconciles every client's bank account, with an agent that writes and runs real Python for each…

Letting an AI agent run code from a random bug report is a security problem. The report is untrusted, the code the agent writes is untrusted, and so is the…

Bank payment migrations fail in public and expensive ways: TSB's 2018 migration cost about £81m in fines and compensation. SwitchProof is an AI test team for…

It analyzes cloud bill, finds the top most expensive services, analyze those services in detail and then come up with actionable insights to reduce those costs.