Pedal

Inspiration

I was sitting in a local coffee shop right across from our university campus, watching the owner stress about the empty tables. Yet, right outside the window, hundreds of students were walking by, glued to their phones. It hit me: The Last Mile Disconnect. Local businesses have products students want, and students have the hyper-local social influence businesses need, but there was no bridge connecting them. The only existing options were expensive agencies or awkward, unscalable manual DMs. I was inspired to build that bridge—a protocol that turns social capital into a tradeable asset.

What it does

Pedal is a dual-sided marketplace that automates word-of-mouth marketing.

For Businesses: It allows local shops to post "Gigs" (e.g., "Post a Story for a free pizza") and hire students instantly without manual negotiation.

For Students: It provides a feed of nearby gigs where they can earn money or perks using their Instagram influence.

The Magic: It features Trustless AI Verification. Students upload a screenshot of their story proof, and our AI agent instantly analyzes it for view counts and tags, releasing payment automatically without human intervention.

How we built it

We leveraged Base 44, an AI-powered full-stack builder, to rapidly prototype a complex architecture:

Backend: We designed a relational database linking Users, Gigs, and Applications with a strict onboarding logic that filters users into "Student" or "Business" roles.

AI Logic: We integrated a Computer Vision LLM to handle the proof verification, turning image data (screenshots) into boolean logic (Pass/Fail) for the payment trigger.

Design: We applied a "High-End Dark SaaS" aesthetic (inspired by Linear and Framer) using glassmorphism and deep void colors to appeal to our Gen Z user base.

Challenges we ran into

The "Gatekeeper" Logic: Ensuring users couldn't bypass the onboarding forms (Identity Proof for students, Shop Photos for businesses) was tricky. We had to implement a strict profile_completed flag that redirects all traffic until the profile is verified.

AI Hallucinations: Getting the AI to reliably distinguish between a "real Instagram story" and a "random screenshot" took multiple iterations of prompt engineering.

Trust: Building a UI that looked professional enough for businesses but cool enough for students required balancing "Fintech Trust" with "Cyber-Streetwear" aesthetics.

Accomplishments that we're proud of

The "No-Touch" Workflow: We successfully built a flow where a gig can be posted, claimed, verified, and paid out without a single human clicking "Approve."

The Design System: We achieved a visually stunning "Dark Apple Liquid Glass" aesthetic that stands out from typical hackathon projects.

Validation: We built a live feedback loop directly into the landing page to capture real user pain points during the demo.

What we learned

The Power of Protocols: We learned that by turning an operational problem (manual verification) into a software problem (AI verification), you can scale a business model that was previously impossible.

Prompt Engineering is Coding: Writing the perfect system instruction for the AI Vision model was just as rigorous as writing a Python function.

User Psychology: We learned that "Trust Signals" (like the secure badge and high-quality UI) are just as important as functionality when asking users for ID proof.

What's next for Pedal

Pedal Pay: Integrating with POS systems so students can scan a QR code to redeem earnings instantly as discounts.

Proprietary Data: Building a "Real-World Influence Score" based on actual footfall conversion, not just follower counts.

Expansion: Launching the "Beachhead Strategy" at our university campus to prove the liquidity model before scaling to other college towns.