How a Hackathon Team Built an AI Emergency Response System in 3 Hours
Chapters7
Announces Sentinel as the global winner of the 11 Labs Worldwide Hackathon and outlines the event scope across 33 cities and 1300 builders.
A 3-hour hackathon win yields Sentinel: an offline-to-online AI emergency-response system that uses 11 Labs’ voice tech to triage crises when networks fail.
Summary
ElevenLabs’ team showcases a remarkable hackathon win by Sentinel, built by Jason Yei’s group to keep emergency lines alive even when internet access collapses. The project leverages 11 Labs’ newest voice models to listen, understand, and coordinate rescues, transforming futile phone calls into structured data packets for responders. Sentinel acts as an offline-to-online bridge, delivering real-time data through voice and scaling to handle surges in non-emergency traffic while preserving fragile networks. The system not only communicates with victims but also automates location extraction and call summarization for quick situational awareness. A live demo places Jason in Greenwich as fire rages, with Sentinel guiding on evacuation steps and identifying the nearest hospital—Queen Elizabeth Hospital—while summarizing results on a dashboard for first responders. The team emphasizes rapid triage: intercept high-volume calls, distinguish true emergencies, and feed verified data to rescue teams. The narrative ties to Australia’s Black Summer fires, illustrating the real stakes when infrastructure fails and calls overwhelm dispatch centers. Jason and teammates—Nin (AI/CIO), Von (CTO), and Andrew (CFO)—present their approach as a mission to ensure families like Sam never wait for help again, culminating in a compelling showcase of how AI can save seconds when every moment counts.
Key Takeaways
- Sentinel uses 11 Labs’ voice model to listen, understand, and coordinate rescue operations during disasters.
- The system acts as an offline-to-online bridge, delivering real-time data through voice even when the internet is down.
- During surge conditions, Sentinel scales to handle non-emergency call volumes with ultra-low latency, reducing bandwidth strain on fragile networks.
- Sentinel automates location gathering and call context summarization, presenting a clean data packet for first responders.
- A live demo places the user in Greenwich with guidance on evacuation and nearest hospital, while a dashboard summarizes map data for responders.
- The team views Sentinel as a first line of defense to intercept mass-volume calls human dispatches can’t handle in time.
- Founding team members are Jason Yei (representing Sentinel), Nin (AI/CIO), Von (CTO), and Andrew (CFO), who assembled the project within months from a prior hackathon.
Who Is This For?
Essential viewing for developers and emergency-response planners interested in AI-driven triage and offline-first communication during disasters. It demonstrates practical use of voice AI to alleviate overwhelmed dispatch centers.
Notable Quotes
"Sentinel is an offline to online bridge that connects victims by delivering live realtime data through voice, even when the internet is down."
—Core value proposition explained during the setup of Sentinel.
"The line is busy right now. The estimated wait time is 15 minutes. Time wasted is lives lost."
—Illustrates the problem Sentinel aims to solve during surge conditions.
"Sentinel to listen, understand, and coordinate rescue."
—Summarizes how Sentinel uses voice AI to triage and guide responders.
"Should I start evacuating? You don't need to evacuate yet. Close all windows and doors."
—Live demo moment showing actionable guidance to a user in danger.
"And with this information, the information is now summarized onto a dashboard provided for first responders to see."
—Describes the operational handoff from victim data to responders.
Questions This Video Answers
- How does Sentinel handle emergency triage when the network is down?
- What makes 11 Labs’ voice model suitable for disaster response demos?
- How can offline-to-online bridging reduce dispatch wait times during crises?
- What are the data privacy considerations when automating victim data in emergencies?
- Which roles contributed to Sentinel (CFO, CIO, CTO) and what does each bring to disaster tech?
11 LabsSentinelAI Emergency Responseoffline-to-online bridgevoice modeldisaster managementemergency triagereal-time data dashboards
Full Transcript
[music] On December 11th, 11 Labs hosted a worldwide hackathon. 1,300 builders came together in 33 different cities to develop projects and ship them in just three hours. In that time, teams created everything from gaming companions to shopping assistants and training programs for employees. Out of all submissions worldwide, one team was selected as the winner. Please welcome Jason Yei representing Sentinel, the global winner of the 11 Labs Worldwide Hackathon. [music] In 2019, Australia experienced one of its worst natural disasters in history, the Black Summer Bush fires. It burned over 24 million hectares of land, an area equivalent to the size of the entire entirely of United Kingdom.
Imagine driving from London to Edinburgh and seeing nothing but ashes. And worst of all, 33 people lost their lives, many waiting for help that never came. Meet Sam. Sam is a father of two and he looks outside and he sees a glowing red sky. He prepares to evacuate his family. But when he turns to his phone for information, the screen is blank. Internet infrastructure is down. There's no maps, no updates, no escape, and no escape routes as well. And Sam's not alone. Thousands of families are also calling simultaneously panicking, asking for simple advice like, "Is the road open?" overloading the ready fragile network.
This exacerbates an 8,400% surge in call volume. And in just 24 hours alone, traffic explodes from 164 calls to over 14,000, completely burying human dispatches. The situation became so critical that the state emergency surfers were forced to shut down the systems just because it couldn't handle the core surge. And when Sam desperately dials for help, this is what he hears. The line is busy right now. The estimated wait time is 15 minutes. Time wasted is lives lost. Every second counts. And that's why we built Sentinel, the lifeline that never hangs up. Sentinel is an offline to online bridge that connects victims by delivering live realtime data through voice, even when the internet is down.
And when a network spikes, it scales instantly to answer non-emergency call volumes with ultra low latency. And with phone calls using thousand times less bandwidth than the internet, it actively preserves the already fragile network infrastructure. And Sensel doesn't just talk. It supports rescue teams by automating location and victim data and summarizing the call context or nearly in a summarized dashboard. By using the latest 11 Labs voice model, we developed Sentinel to listen, understand, and coordinate rescue. First, it intercepts when disaster strike and cause spike. Sentinel is the first line of defense, intercepting mass volume calls that human dispatches cannot physically handle.
It then listen provides guidance and feeds live datas to victim filtering filtering out background noise and panic to ex uh extract verified data packets and most importantly providing location and summarizing the data of the victims. Afterwards, a trigger is a call conducting real time risk analysis to distinguish between questions like, "Is the road open to I'm in a trapped in a burning house? Please help me." Escalating the true emergencies. And finally, rescue. It presents a verified data packet for a clean dashboard for first responders to see to make sure every second counts. And now, let's go to our live demo.
How can I help? My name is Jason and I'm currently at the Magazine London Greenwich. There's fire on the horizon and I'm worried that it's bush fire. What should I do? Are you in immediate danger right now? I'm not in danger right now, but there's smoke on the horizon and it's coming towards my direction. Should I start evacuating? You don't need to evacuate yet. Close all windows and doors. Stay indoors to avoid smoke exposure and be ready to leave if the smoke gets closer or authorities give instructions. I'm here if anything changes. And just in case, where's the nearest hospital?
Queen Elizabeth Hospital is the nearest major hospital to you in Greenwich. And with this information, the information is now summarized onto a dashboard provided for first responders to see with the live map data as well. And please meet the team that brought Sentinel alive. Uh so we we only met four months ago at a different hackathon and we have Andrew who's our CFO specializing financial software and also Nin our CIO who's an AI specialist and Von our CTO who's focusing on building for an impact. Thank you very much and join us today and ensure that families like Sam never have to wait for help again.
Thank you. [applause]
More from ElevenLabs
Get daily recaps from
ElevenLabs
AI-powered summaries delivered to your inbox. Save hours every week while staying fully informed.






![[VOD] Laravel AI SDK just dropped (part 1) thumbnail](https://rewiz.app/images?url=https://i.ytimg.com/vi/nvhdOUiYAI4/maxresdefault.jpg)