Skip to content

AI lab hackathons, 2025 to now

What got built, and what won.

Every submission from hackathons run by Anthropic, OpenAI, Google, Meta and the other big labs, with the winners marked and the code linked.

projects
31,980
hackathons
93
award winners
585
inferred from GitHub
433
Arize - 1st Place

Cassandra

AI agents fail silently, confidently wrong, and nobody notices. Cassandra is an AI that watches your AIs. It catches every lie, diagnoses why, ships a proven fix, and even grades its own accuracy.

agentsdev tools

Sirjan Singh, Kshitij Verma

Fivetran - 1st Place

Unravel

Five agents watch evolving variant evidence, re-score a clinic's uncertain DNA findings with a calibrated Bayesian model, and draft family recontact the moment one turns actionable. Draft-only.

healthscience
GitLab - 1st Place

mnemosyne

The AI that remembers every failure — predicts outages, annotates risky code inline, gates CI pipelines, auto-blocks critical MRs, and diagnoses failures. Powered by Gemini and gitlab on Google Cloud.

dev toolssecurity
MongoDB - 2nd Place

KickOff

Seven Gemini 3 agents move 82,500 World Cup fans to MetLife Stadium — reasoning over live MongoDB Atlas via MCP, every decision a database write streamed onto a photorealistic 3D map.

agentsproductivity
Arize - 3rd Place

ClinTrace

ClinTrace is an AI agent that assists in emergency department triage — severity scoring, red-flag detection, specialist routing — with every reasoning step traced and auditable via Arize Phoenix.

healthagents
Elastic - 3rd Place

Carryover

Visa intelligence that protects migrants from agency fraud. Three AI agents on Elasticsearch + Gemini that don't just inform but they detect scams in real time and file real complaints.

securityagents
Fivetran - 3rd Place

How Did I?

How Did I? is an agentic ecommerce operator that diagnoses daily revenue leaks across carts, ads, inventory, and support—then drafts owner-approved fixes in one tap.

financeproductivityagents

Cova

NFC fitness ring that bridges physical and online stores through fitness-based rewards. Tap to connect with friends, share workout metrics, earn rewards, and eventually pay with your ring.

robotics/hardwarehealth

FRAUDINTEL

Stop reacting to fraud. FraudIntel is a proactive AI agent that autonomously investigates threats, connects the dots across your data, and executes real-time countermeasures.

securityagents

MarginTrust AI

An AI revenue leakage agent that uses Gemini, Google Cloud, and Fivetran to detect stale or broken SaaS data, estimate dollar impact, and recommend what Finance, RevOps, or Data teams should fix first

financedev tools

Nikhita Shankar, Manas Tokale

Gauge

Merge requests and CI for knitting patterns. A robot test-knits every row of every size in software and fails at the exact row where the stitch math breaks — before anyone wastes yarn.

creative/musicdev tools

Graft

Fivetran moves 750+ sources. Your obscure tool isn't one of them — so Graft's agent writes, tests, deploys, and runs a real connector for it, live, while you watch. No connector? Graft one.

dev toolsagents

Murmur

Every voice a bird. Murmur turns thousands of raw feedback lines into a living murmuration — ask anything and Atlas $vectorSearch re-flocks the sky into the answer, live, in seconds.

productivity

SLEUTH

SLEUTH is an autonomous incident investigation agent using Gemini AI + Google Cloud Agent Builder + Elasticsearch to investigate incidents by correlating logs, traces, and metrics in real-time.

dev toolsagents

Field Force

An autonomous AI dispatch crew that turns a field-service emergency into a ready-to-approve action plan in seconds, built on Gemini 3 and Elastic's MCP server.

agentsproductivity

RepairBox

An autonomous agent safety engineer that senses agent failures via Arize Phoenix, synthesizes safe patches, replays comparisons, and gates deployments with before/after evidence.

dev toolsagents

SyncMind

SyncMind is an autonomous agent that keeps your business documentation synchronized with your codebase in real time, so the gap between what you planned and what you shipped never has to exist again.

dev toolsproductivity

Al-Dia Expense Agent

Agente de IA para finanzas familiares. Registrá gastos por chat, audio, foto o PDF y un equipo multi-agente con Gemini (ADK + MongoDB MCP) los convierte en dashboards, análisis y Excel.

financeagents

Mateo Camelino, Lisandro Monti, Ignacio Marinelli +1

Skywave

Type your callsign and watch a live map of everywhere on Earth your radio signal is reaching right now — then ask when you'll reach Japan, answered from what actually happened on the air.

science

Harmony

An HS-code classifier that publishes its own error rate. Graded nightly on federal rulings it has never seen, it reads its misses over Phoenix MCP, patches itself — and ships only if the score rises.

legalagents

Kineticz

Autonomous schema-drift remediation that draws on its own repair history and lands every fix as a reviewed, tamper-evident GitLab merge request.

dev toolsagents

Darn.

When a deploy tears a hole in production, Darn finds the exact commit, writes the fix, and opens a PR where every claim is a receipt you can re-run in Dynatrace. A human approves. Davis confirms.

dev tools

payLens

PayLens: two AI agents for payments. Dev Studio tests Stripe/Paystack in plain English. Recon Studio auto-matches invoices to payments. Powered by Gemini, Elastic, and Google Cloud.

financedev tools

toluwani Bakare, Emmanuel Elijah, Olamide Olanrewaju

ProofPilot

ProofPilot uses Google ADK & Elastic to turn API docs & customer context into functional, source-grounded React+Node demos in seconds, with built-in claim checking to avoid hallucinations.

dev toolsagents

Sameer Dhole, MaxPollardII Pollard, Kevin Wang

MedCred

AI health content hallucinates up to 48% of the time. MedCred is the integrity layer that catches failures before they reach patients - and gets smarter after every eval.

healthdev tools

EcoOps

Making autonomous AI sustainable. EcoOps acts as a digital manager that instantly breaks AI hallucination loops, saving both your cloud compute budget and planetary carbon.

dev toolsscience

Crucible

One database in → one tuned, measured text-to-SQL agent out. Crucible reads its own Arize Phoenix traces via MCP, hypothesizes the dominant failure, mutates, and re-scores until it wins.

dev toolsagents

Mission Control

An AI life coach built on Google Agent Builder & Gemini. It transforms chaotic brain dumps into load-balanced tasks across your defined life pillars, saving directly via the MongoDB MCP.

productivityagents

Neural 360

Neural 360 unifies every customer signal—sales, support, usage, sentiment—into one unified profile, turning scattered data into real-time, explainable recommendations that grow revenue and cut churn

productivity

Mrudhula Katta, Kalpana Dakoji, Hansika Reddy