The Zerops Challenge

The Zerops Challenge

Break free from infrastructure drag

Build a working product, deploy it on Zerops, and take it from an idea to a live application in one weekend.

Prizes

MacBook Neo, Logitech MX Master 3 & $5,000 in Zerops credits

Registration

Loading…

One moment. Pulling up this hackathon.

When
August 8–9, 2026
Where
Online, from anywhere
Format
Solo only, one project each

Winners

Who took the prizes home

The best project and the best build post, then the standouts who split the Zerops credits pool.

  1. Best overall project

    Winner

    Deplot AI

    Autonomous Platform Engineer for Zerops

    Manan · @maannaan

    Deplot is an AI platform engineer for Zerops: paste any GitHub repo and it analyzes the stack, designs architecture, generates zerops.yaml, deploys into an isolated sandbox, then operates production with live health, AIOps diagnosis, self-heal (env patch + redeploy), and a deployment readiness score. Unlike config-only checkers, Deplot ships and operates on Zerops — repository intelligence → architecture → deploy → observe → heal. Built with Next.js + FastAPI, specialized LLM agents, and real Zerops/zcli integration.

    The MacBook Neo awarded for the best overall project

    Wins

    MacBook Neo

  2. Top social post

    Winner

    FlowForge AI

    Autonomous Site Reliability incident responder

    Raj Tiwari · @rajtiwari

    FlowForge AI is an autonomous, production-grade Site Reliability Engineering (SRE) incident triaging and diagnostic platform. Instead of developers wasting hours parsing log stacks and checking database connection limits during critical outages, FlowForge AI serves as an active, on-call virtual SRE that reduces Mean Time To Resolution (MTTR) from hours to under 2 seconds. Core Features: 1. Telemetry Incident Simulator: Inject real-time PostgreSQL connection slot exhaustions, Valkey memory leaks, or CORS Gateway handshake timeouts. 2. AI Multi-Agent Storytelling Pipeline: Visualizes a collaborative graph of 4 specialized agents (Observation, Diagnosis, Resolution, and Infrastructure) executing live diagnostics. 3.Codebase Remediation Patches: Generates exact code patch blocks (TS / Config) to fix the isolated codebase bugs. 4.Zerops Infrastructure Advisor: Generates horizontal/vertical autoscaling configurations and rationales mapped to the service topology. 5. SRE Dave Interactive Chat: A persistent SSE-powered chatbot to query follow-up diagnostics. 6. Live Topology Map: Visualizes network packet flows running across services. How it uses Zerops: The monorepo deploys 5 isolated microservice containers (React static client, Fastify API, BullMQ worker, Valkey queue, and PostgreSQL) orchestrated entirely on Zerops PaaS. We utilize Zerops' private WireGuard mesh network to block database access from public internet ports, dynamic env references to automate database bindings, and static L7 load balancing to route traffic seamlessly.

    The Logitech MX Master 3 awarded for the top social post

    Wins

    Logitech MX Master 3

Zerops credits

5 winners
  1. ZeroArch

    Visual AI Infrastructure Designer

    Cloud infrastructure shouldn't require a Ph.D. in YAML. ZeroArch is a visual, AI-powered infrastructure designer built specifically for Zerops. Instead of writing error-prone deployment scripts by hand, developers can drag and drop scalable components (React, Node.js, PostgreSQL, Redis) onto an infinite React Flow canvas. When the design is complete, a custom Google Gemini AI pipeline acts as a DevOps engineer—auditing the architecture for security flaws and compiling a production-ready zerops.yml configuration instantly. Architectures are persisted to PostgreSQL, allowing users to generate shareable, read-only links to collaborate with their teams. Built with Next.js 15, React Flow, Prisma, and Google Gemini AI, and deployed natively on Zerops.

  2. ScaleScope

    Prince Panchani · @princehunter

    A live autoscaling proving ground, built on and about Zerops. ScaleScope fires controlled load at a service and shows Zerops autoscaling in real time — container count, latency, and throughput streaming live, with every run replayable afterwards, a digital twin predicting what happens next, and unattended experiment suites that turn one run into a swept curve.

  3. Ephemera gives every pull request its own full stack. Open a PR and you get an API, a database, and a live URL. Close it and the whole thing disappears. Frontend teams have had this for years with Vercel and Netlify. Backend teams haven't. The full stack options I found (Bunnyshell, Northflank, Qovery) assume Kubernetes and sell upmarket, so small teams and open source maintainers running docker compose are left sharing one staging database and queueing up behind each other. Ephemera is preview environments for the people who are never going to run Kubernetes. One preview.yml, ten minutes, real infrastructure. The architecture is the part I care most about. The GitHub webhook verifies its HMAC signature, writes down what state the environment should be in, and returns in about 40 milliseconds. That's all it does. A separate worker then reconciles what should be true against what is true, taking exactly one idempotent step per tick: pending, provisioning, deploying, ready, destroying. Provisioning infrastructure is slow and fails halfway through, so every step has to be resumable. I killed the worker mid provision over and over while building this; it always converged instead of leaving half created environments lying around. The control plane itself runs on Zerops as five services: a static React dashboard, a Hono API, a private worker, Postgres, and Valkey. Preview environments land in a second Zerops project, created with zcli service import, deployed with zcli push, and exposed on generated subdomains. A cron reaper destroys anything past its TTL, because a forgotten preview environment is just a bill you haven't seen yet. A few things I'm glad I got right. An environment is only marked ready once an HTTP request to its public URL actually succeeds, because "the container exists" and "the app works" are very different claims. Readiness survives a transient 502 instead of giving up on one bad poll. Concurrent provisioning waits for Zerops' per service process queue to drain before enabling subdomains, which was the last real bug I chased down. Dashboard mutations need a bearer token while viewing stays public. And the provider sits behind a four method interface, so Zerops is the first adapter rather than something the project is welded to. Measured end to end: pull request opened to live URL in roughly 100 seconds. There's also a docker compose to preview.yml importer, because setup friction is the real reason small teams don't already have preview environments. Demo video: https://youtu.be/bmVL_aqa034 Judges: viewing the dashboard is public, but the destroy, retry, extend TTL, and run demo buttons need this admin token: f2d58a1908f4d0e802a538e3cecf8b5f2c008a343fe7d73fe325454e2c33316f Paste it into the Admin token field at the top of the dashboard. If the live demo environment has expired by the time you look, press Run live demo and a fresh one comes up in about 90 seconds. The repo being previewed is github.com/GautamTalksDev/ephemera-demo-app. AI tools used: Cursor with Claude for writing code, and Claude for architecture and debugging. The design calls and the platform level diagnoses were mine: the Postgres minimum disk floor, zcli needing the noGit flag, Zerops running start commands through exec so shell operators don't work, cross service environment variable translation, and the subdomain process queue race. The commit history shows how that went.

  4. DeployDoctor

    AI-Powered Deployment Incident Diagnosis

    DeployDoctor turns deployment and runtime logs into actionable incident diagnoses. It uses deterministic analysis for known failure patterns and Google Gemini for unfamiliar or low-confidence incidents, returning the likely cause, severity, evidence, recommended fix, and next steps. Built with React, Fastify, PostgreSQL, Prisma, and Google Gemini, DeployDoctor is deployed on Zerops as a multi-service application with GitHub-based CI/CD.

  5. Doorbell

    Doorbell - a self-hosted tunnel that doesn't lose your webhooks. Close your laptop and ngrok, Cloudflare Tunnel and localtunnel all hand your senders a 502. None of them can do better: holding your payloads would mean keeping your signatures and tokens on their servers. Doorbell holds them because the database is yours - a request for an offline tunnel is written to Postgres before the sender is answered, so it gets 202 Accepted instead, and the queue drains in order the moment you reconnect. That is only possible on Zerops. The CLI holds one raw public TCP port open and multiplexes every request back down it. It isn't HTTP, so no serverless platform can express it - a function stops when it answers, and nothing is left holding the socket. Zerops gives you that port, a process that never sleeps, and managed Postgres and Valkey on a private network, from one pasted YAML file. Three services, ninety seconds. The platform choice isn't a preference here, it's the feature. More details on readme - run locally or use curls Live: https://gw-2ad0-3000.prg1.zerops.app Dashboard: https://gw-2ad0-3000.prg1.zerops.app/dashboard?token=yg_0Gg22aYNgH4LTZL_ZxBC4evSIk7pF Video URL: https://youtu.be/KIYEptODPeo X Post: https://x.com/big_achieverx/status/2086540506033709544?s=20

Projects

271 projects, one weekend

Every submission to The Zerops Challenge, in the order it came in. Each one was built by a single hacker and deployed on Zerops, so each one has somewhere to go and look: the running deployment, the source it was built from, and the post its author wrote about it where there was one. Links are as they were submitted, so a deployment that has since been taken down may no longer answer.

271 projects