AgentMail to Host 24-Hour AI Agent Hackathon at YC Office with $1,000 Costume Contest

San Francisco – AgentMail, a Y Combinator Summer 2025 (YC S25) startup, is set to host a 24-hour Halloween-themed hackathon, "HackHalloween," at the YC office starting at 6 PM on October 31. The event, focused on building autonomous AI agents, is sponsored by leading tech companies including Replit, Browserless, LiveKit, Convex, and Hyperspell. Participants will compete for tens of thousands of dollars in prizes, with winning teams guaranteed interviews for the prestigious YC program.

The hackathon emphasizes the development of AI agents, reflecting a growing trend in the tech industry. AgentMail, which provides API-first email services specifically for AI agents, aims to foster innovation in this rapidly evolving field. Cofounder Ankit Gupta expressed excitement on social media, stating, "Super excited that @agentmail is hosting a spooky Halloween themed Hackathon... at the YC office starting at 6pm on Oct 31 and running for 24 hours!"

Beyond the technical challenge, "HackHalloween" will also feature a unique $1,000 costume contest, encouraging a festive atmosphere alongside intensive coding. The event is limited to 200 builders, highlighting its exclusive and competitive nature. Y Combinator's involvement, including providing the venue and offering interviews to winners, underscores the accelerator's commitment to nurturing next-generation AI talent.

AgentMail, founded by Haakam Aujla, Michael Kim, and Adi Singh, offers a specialized email infrastructure that enables AI agents to send, receive, and act upon emails, addressing limitations found in traditional email providers. The company's participation in the YC S25 batch positions it at the forefront of AI agent development, with its platform designed to facilitate programmatic inbox creation, usage-based pricing, and real-time event handling for AI systems. This hackathon serves as a significant platform for AgentMail to engage with the developer community and showcase the potential of agent-based applications.