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
Student
Photograph a plant. Find out if it belongs there, and what to do about it.
Featherless serves 21,702 open models. Model Scout tells you which one to actually use — it runs a real bake-off on your task and ranks them on quality, latency and cost.
Every agent with an inbox can be hijacked by the inbox. Airlock asks permission on a channel the attacker can't reach. The second channel isn't a notification pipe, it's the boundary.
Scam blockers work after the call. Kavach works during it - warning you live, in plain words, why a call is a scam and whether the voice is an AI clone. Runs on your device.
Every grievance in India gets a number. Far fewer get a hearing.
Catch the upstream change before the model fails. An agent that traces DataHub's ML lineage to find schema changes that silently break production models - and writes them back as Incidents.
Governed AI creative: generate ad campaigns that can prove where they came from.
KitabDost turns one English textbook page into a Hindi teach-back and mastery card, helping class 6–10 learners show what they truly understand.
Nonprofits live in Slack but coordinate volunteers by phone tree and spreadsheet. Rally fills shifts by skill, auto-rescues cancellations, and answers volunteer questions with sources.
ChainSentinel forecasts material delays with foundation models, then an agent pulls the contract, alerts the right people, and remembers how every disruption was fixed. Built for construction.
Mnemosyne is the first sparse-memory substrate that lets language models learn continuously without retraining — replacing dense attention with a biologically-plausible consolidation loop.
