Devpost
Participate in our public hackathons
Devpost for Teams
Access your company's private hackathons
Grow your developer ecosystem and promote your platform
Drive innovation, collaboration, and retention within your organization
By use case
Blog
Insights into hackathon planning and participation
Customer stories
Inspiration from peers and other industry leaders
Planning guides
Best practices for planning online and in-person hackathons
Webinars & events
Upcoming events and on-demand recordings
Help desk
Common questions and support documentation
Sort by:
A shop's own camera becomes its checkout and stock clerk: NVIDIA Nemotron on Nebius Token Factory decides what was taken, names it, advises at the exit and answers the owner.
One iPhone LiDAR walk turns your home into a digital twin any robot can find itself in. Say "check the kitchen" in any language: Nemotron plans it, and the robot asks when unsure.
Draw a game on a napkin. An AI studio of NVIDIA Nemotron agents builds it in 3D, playtests it with vision, fixes what isn't fun, and ships a link you can race your friends on.
Eight taps plan your night out, powered by Qloo. Your agent holds it all, and when a booking page tries to trick it, your taste twin catches it. Your night, your taste. No evil twins.
Customers chat on WhatsApp; an AI agent builds a checked cake order, quotes it, checks the bakery’s calendar, and staff approve what matters. Confirmations by WhatsApp and SES, on Bedrock AgentCore.
Autonomous multi-agent collaborative visual SLAM and 3D reconstruction for communication-constrained disaster environments.
Four lost hikers, one hunting predator. OpenCV 5 cameras feed a shared living map. Across 160,000 pre-registered simulations, sharing what each hiker sees raises escapes from 27% to 42%.
Faster vision models can silently miss objects. VisionProof tests each OpenCV speedup on labeled images, shows every lost detection, and blocks regressions before they ship.
Instead of guessing whether furniture fits through a door way or if there is enough room for everything you can use RoomTwin to be sure. Paste a link from IKEA or any furniture and see.
An agent that speeds up OpenCV 5 pipelines and refuses any speedup that quietly breaks the output.
Turn ordinary multi-camera videos into animation-ready 3D motion—no suits, markers, or motion-capture studio required.
Ask your video where things are. Every answer comes with the frame that proves it.
Jharia's coal seams have burned since 1916. We track every fire from Landsat with OpenCV 5, project its spread 48 days out, and an agent closes threatened roads. A person reopens them.
The problem: Small e-commerce businesses pack orders manually. Missing items or incorrect product variants lead to extra shipping costs, returns, and unhappy customers.
Voice guidance for blind visitors in any building, after two volunteers walk it with their phones for 15 minutes. No beacons; 1.4 cm on benchmark video.
Tell a camera its job in Uzbek, Russian or English. Ko'z builds it from OpenCV 5 blocks, tests and fixes itself on live frames, blurs faces, and stops when it sees a toilet or bed.
A bird and a drone look the same at 30 fps. Only the drone hums. Event cameras + OpenCV 5 on COOL/Graviton4 confirm drones by propeller hum: 25% fewer false alarms, same recall.
Help small workshops without vibration sensors spot changes in machine vibration using a slow-motion phone video—and know when a closer inspection is needed.
A camera that forecasts landslides: OpenCV 5 measures ground creep to 0.005 px, forecasts the failure window, an AI agent re-checks it, and a person decides.
SLAM that knows when it is lying: an OpenCV 5 agent diagnoses tracking failures, applies one bounded fix, verifies it, and rolls it back if it did not help. Every decision is audited.
Clouds shouldn’t end your investigation. ClearSky’s OpenCV 5 agent finds before-and-after satellite images, checks alignment, and helps recover failed comparisons through evidence and human approval.
A debugger for the physical world: OpenCV 5 and AWS turn camera observations into actions, learn from missed attempts, and plan better moves.
Crack width in millimetres from a phone photo, with a calibrated gate that refuses unreliable readings and an agent that tells you how to retake the photo.
Floor posture safety watch
