Livemap: Context-Aware Idle Capacity Matcher

A hyper-local marketplace that matches a user's travel "wait time" with a merchant's "live idle capacity," featuring AI intent matching and a "Time-Safe" return-to-station navigation.

Comment

Livemap: Turning Travel "Dead Time" into Wellness Windows

💡 Inspiration Have you ever waited at Zurich Airport or SBB with 100 minutes to kill, just doomscrolling on your phone? Meanwhile, a 5-minute walk away, a local massage studio has an empty chair because of a last-minute cancellation.

Why don't we connect them? Because of Time Anxiety. People don't book services during layovers because of the "fear of missing the train." Existing map apps solve the spatial problem (where things are), but ignore the temporal reality (when you need to be back).

Our inspiration was to eliminate this anxiety entirely. By making Time the core metric of our marketplace, we align perfectly with Livemap's vision of true context-aware matching: connecting users with what they need, exactly when they have the time for it.

⚙️ What it does We built LiveM, an end-to-end platform that turns fragmented waiting time into actionable demand, seamlessly matching it with local idle capacity:

For the User (The Demand):

Automated "Relax Window" Calculation: A user inputs their trip (e.g., SBB IC 8). The system instantly calculates the layover (e.g., 101 mins) and secures a safe "Relax Window" (e.g., 76 mins), factoring in transit buffers.

Multi-Turn AI Intent Parsing: Instead of complex filters, users speak naturally ("I'm tired, I have 100 mins"). Powered by Featherless, our AI maps this to nearby services. Users can even refine results on the fly ("Please find something cheaper!"), and the AI instantly adapts the map.

Transparent "Time-Safe" Guarantee: Before booking, the user sees a crystal-clear visual breakdown: Walk Time + Service Time + Safety Buffer = Total. The green "You're safe!" badge provides ultimate peace of mind.

The Return Alert: Once the session is running, a hard pop-up interrupts the user when it's exactly time to leave ("Time to head back. Train departs in 20 mins"), triggering seamless return navigation.

For the Business (The Supply):

3-Click Live Broadcast: A hyper-fast dashboard for busy shop owners. Spot an idle gap? Click the slot, set a dynamic discount (e.g., 30% off), and broadcast to the LiveMap radar in seconds.

Live Client Tracking (Anti-No-Show): Once a match is made, the merchant sees the client approaching on a live map with real-time ETA updates (3 Mins -> 1 Min -> At The Door), entirely eliminating operational anxiety.

🛠 How we built it Knowing that Livemap evaluates practical workflow and high-quality UX over raw backend infrastructure at this stage, we adopted a Product-Led Validation Architecture. We focused our technical bandwidth on the hardest parts: Natural Language Intent Parsing and the End-to-End User Flow.

We integrated Featherless as our core AI reasoning engine to handle real-time, multi-turn natural language processing, turning messy human input into strict JSON queries for routing and booking.

To bring this logic to life, we coupled our AI brain with a breathtakingly high-fidelity frontend generated via Lovable. While the SBB train schedules and map routing currently utilize simulated local states to prove the spatial-temporal matching concept, the core AI-to-Matching translation is fully functional and driven by live LLM inference. This strategic choice allowed us to dedicate 100% of our hackathon time to perfecting the UX, the visual time-breakdown logic, and the frictionless merchant dashboard.

🚧 Challenges we ran into Translating Vague Language into Strict Time Boundaries: Parsing a prompt like "I'm exhausted" into a search for a 60-minute massage, and then dynamically filtering the map when the user says "find something cheaper", required careful prompt engineering and fast inference via Featherless.

Designing the "Trust" UI: It was challenging to figure out how to present complex time constraints without stressing the user out. Designing the intuitive Walk + Service + Buffer timeline card was crucial to making the app feel like a trustworthy companion rather than a strict alarm clock.

The Cold Start Merchant UX: Traditional CRM booking systems are too slow for spontaneous "live drops." Our biggest supply-side challenge was condensing the broadcast flow into just 3 clicks, while providing them with the live-tracking map to ensure they actually trust the system.

🏆 Accomplishments that we're proud of The Ultimate Context-Aware User Flow: We successfully engineered a flawless, highly visual user journey that perfectly bridges a user's train schedule, their emotional intent, and a merchant's live availability.

The Multi-Turn AI Experience: Implementing a conversational layer that actually feels helpful and dynamically updates the map view based on follow-up requests.

Frictionless Dual-Sided Trust: We solved anxiety for both sides: the "Head back" alert keeps the user safe, while the "Live Client Tracking" map keeps the merchant confident that the customer is actually arriving.

🧠 What we learned We learned that in ultra-local marketplaces, Context > Discovery. Users don't just need to see a map of what's around them; they need the system to synthesize their personal schedule, walking speed, and the merchant's exact real-time availability. We also validated that an impeccable, anxiety-free workflow is the strongest way to prove a product's market fit.

🚀 What's next for LiveM Live API Integrations: Swapping out synthetic state data for real APIs (Google Maps/Mapbox for live routing, SBB Open Data API for real-time train delays).

Real-Time WebSocket Backend: Building the lightweight infrastructure to handle the actual live synchronization between the User app's location and the Merchant's tracking radar.

Pilot Testing: Onboarding 5-10 local businesses near Zurich HB or Kloten to test the "Live Drops" feature with real layover foot traffic.

💻 Built With Featherless (Real-time LLM Inference & Multi-turn Intent Parsing)

Lovable (Rapid UI/UX Prototyping & React Code Generation)

React / TypeScript / Tailwind CSS (Frontend Stack)

Simulated State Management (For executing the Spatial-Temporal matching logic)

Synthetic-Data State Management (For conceptualizing the Spatial-Temporal matching logic)

AI Prompt Engineering

💡 Final Note: A Zero-Anxiety Marketplace LiveM proves that the biggest barrier to spontaneous local consumption isn't budget or lack of interest—it's the friction of time. By designing a zero-anxiety workflow for the user and a 3-click broadcast tool for the merchant, we’ve built the blueprint for a hyper-local economy that thrives on fragmented time. This is how context-aware matching converts passive waiting into active revenue.

Built With

Share this project:

Updates