vøiddo on devpost · updated 2026-05-16

we ship small working things on real deadlines.

ten active builds for cash-prize hackathons. each carries a public app or a public repo, screenshots, a demo, and an honest scope. judging happens on the work, not the pitch.

building   tracelock · find evil building   tracelock ops · splunk building   launchroom · rapid agent building   edubridge · gemma 4 good building   signal desk · web data unlocked building · deadline jun 29   claudeflow · uipath agenthack building   queue sentinel · microsoft agent academy building · deadline may 22   sentinel · ibm bob lablab.ai

01 active builds

eleven projects, eleven deadlines.

each card links to the local proof page (screenshots, demo, audit notes) and to the hackathon overview on devpost.com. sorted by deadline — closest first.

dupe-detector

reddit mod tools hackathon

a reddit-native duplicate-review mod tool. posters precheck before submitting; mods get a last-7d duplicates queue with similarity scores and one-click lock/remove. zero llm, zero external network. already live on reddit developer platform — only the submission needs to go in.

platformreddit devvit
deadlinemay 27, 2026
app statuslive & approved
external apinone · zero deps

tells lite

build with medo hackathon

a no-code message coach built end-to-end inside medo. paste a thread, get a calmer rewrite, see the rule. tells lite carries the public app, support page, and demo reel.

prize$50,000 cash
deadlinemay 20, 2026
winnersjun 7, 2026
code-freerequired

tracelock

find evil! hackathon

building

a defensive incident agent that refuses unsupported claims. reads case events, forms findings, runs a correction pass, and ties every conclusion to exact evidence. precision and recall stay visible.

prize pool$22,000+
deadlinejun 15, 2026
harnessprecision 1.0
guardrailno write tools

launchroom

google cloud rapid agent

building

a launch-readiness agent for the hour before shipping. inspects release evidence, calls tools, ranks blockers, and returns owner-safe next actions instead of a vague green light. if a required mcp is blocked, the report says blocked.

prize pool$60,000 cash
deadlinejun 11, 2026
stackgemini · mcp
evalstrict, no fake green

tracelock ops

splunk agentic ops hackathon

building

an evidence-first incident triage room for Splunk teams. takes an alert bundle, forms evidence-backed findings, marks every unverified conclusion, and generates explicit next SPL queries — tied to the Security track and Best Use of Splunk MCP Server.

prize pool$20,000 cash
opensmay 18, 2026
deadlinejun 15, 2026
mcp prize$1,000

faultline

devnetwork ai + ml 2026

a resilience console for autonomous operators. shows an agent choosing a safe failure mode instead of pretending success. live TrueFoundry AI Gateway integration — gemini-2.5-flash-lite primary, gemma-3-27b-it free fallback — with 2× authenticated HTTP 200 gateway calls as proof.

prize$1,500 cash
deadlinemay 28, 2026
stacktruefoundry · gemini
gatewaylive + fixture

edubridge

kaggle × google deepmind · gemma 4 good

building

an AI-powered english activity generator for low-resource classrooms. teachers choose an age group and topic; gemma 4 returns a structured, print-ready lesson plan. no internet in class — activities are cached for fully offline use.

platformkaggle tpu v3
trackeducation
offlinepwa · service worker
costzero cloud cost

signal desk

bright data · web data unlocked

building

a live-web research desk for outbound and partner scouting. deterministic by design: fixture mode is fully explorable today; live mode stays blocked until a real authenticated bright data capture is attached. every claim in the brief points back to a retrieval step.

data pathweb scraper api
modefixture · live blocked
evidenceproof-linked brief
use caselead qualification

claudeflow

uipath agenthack 2026

building

ai-augmented bpmn process orchestration: claude classifies contracts, detects exceptions at each uipath maestro stage, and decides retry / escalate / reroute at runtime. no hand-coded catch blocks — the ai is the exception handler.

prize pool$48,000+
deadlinejun 29, 2026
trackmaestro bpmn
bonusclaude api ✓

queue sentinel

microsoft agent academy · special ops

building

a deterministic ops triage agent for the special ops track. reads live vøiddo state and returns a clean run-now / repair / defer / owner-only split — no guessing, no silent failures. copilot studio integration is fixture-safe now; live microsoft path is gated behind one authenticated capture.

sponsor techcopilot studio
trackspecial ops
fixture modelive · explorable
ms proofblocked → pending

sentinel

ibm bob hackathon · lablab.ai

building · may 22

an enterprise AI resilience console built on IBM watsonx. shows how granite-powered agents handle IAM token failures, model brownouts, rate-limit storms, and response validation failures — safely, without silent data loss. fixture mode is fully explorable; IBM watsonx mode unlocks with a real IBM Cloud request ID proof.

sponsor techIBM watsonx
deadlinemay 22, 2026
fixture modelive · explorable
prize pool$10,000

02 how this surface works

proof-not-pitch.

each linked project carries a runnable slice or a public proof page before the devpost submission goes in. no mocks, no fake demos, no invented metrics.

scope rule

each entry on this page maps to an active hackathon with a confirmed deadline and a confirmed prize. when one closes, we move it to a results page — we do not retroactively claim wins.

media

screenshots and demo videos are produced from real runs, not mockups. when a clip is a scripted browser capture rather than a narrated demo, the page says so on the same surface.

cross-link

this hub also lives one click from /reddit/ (our devvit apps surface), the root portal, and the per-project proof pages above.

future builds

scout output is tracked locally and seeded into this board as new active hackathons close in on a deadline. eazo and others are queued but not surfaced until a working slice exists.