Engineering Boksi: Creator-Brand Marketplace, Dual Dashboards, Platform Payments & Gemini AI Content Screening
“Parallel Loop built Boksi from zero: brands filter creators in seconds, orders flow through both dashboards, payments stay on-platform, and Gemini catches AI-fake deliverables before we release funds.”
Boksi is a marketplace where artists and creators produce branded content and get paid. Creators build profiles that showcase videos, social links, and engagement signals so brands can judge reach before hiring. Brands filter talent by age, country, gender, and content niche, then run campaigns with all payments handled inside the platform.
The Marketplace Challenge
Parallel Loop built Boksi from scratch as a two-sided product that must convert visitors, expose creator ROI data, and protect brands from low-quality or AI-generated deliverables. Influencer marketplaces fail when discovery is opaque, order state is tracked in spreadsheets, or payouts happen off-platform. Boksi needed a user-friendly public landing experience, searchable creator profiles, separate creator and brand dashboards, escrow-style payment flows, and Gemini-powered screening on submitted content. Three full-stack developers delivered the React and Next.js front ends, Node.js and Express APIs, PostgreSQL models, and AWS EC2 hosting.
- DM-Based Hiring Without Structured Profiles
- No Unified Engagement Metrics for Brands
- Payments Outside the Platform
- Manual Order Tracking in Email
- No AI Deliverable Verification
- SEO Landing + Creator Profile Directory
- Brand Filters: Age, Country, Gender, Niche
- Creator Dashboard: Orders, Submit, Get Paid
- Brand Dashboard: Hire, Review, Pay
- Gemini AI Check on Submitted Content
Phase 1: Landing Page, Profiles & Creator Discovery
We developed a user-friendly frontend landing page so creators and brands can sign up and browse creator profiles. The marketing site explains value for both sides, supports authentication, and surfaces a directory of creators with portfolio videos and linked social accounts. Engagement metrics are normalized so brand buyers can compare audiences before sending a brief. PostgreSQL stores profile attributes used later in filters: age band, country, gender, and content niche tags.
Our custom software development team built responsive React components on Next.js with server-rendered profile pages for shareable creator URLs and fast search indexing on boksi.com.
Phase 2: Creator & Brand Dashboards, Payments & Gemini Screening
We developed a creator dashboard where talent lists content, tracks orders, produces brand assets, submits deliverables, and receives payment. Order timelines show brief details, deadlines, revision notes, and payout status. Creators attach final video or image packages directly in the workflow instead of emailing files.
For brands we developed a dashboard to hire creators, review produced content, and complete payments on-platform. Brand users run filtered searches, send hire requests, approve submissions, and release funds when work meets campaign requirements. To ensure deliverables are authentic, we integrated Gemini APIs to detect AI-generated content before approval. Our AI and machine learning engineering layer flags synthetic media so brands do not pay for fraudulent clips.
- Creators: Profiles with videos, social links, and engagement visibility.
- Brands: Filter by age, country, gender, and niche; hire and pay in one place.
- Payments: Platform-handled transactions from hire through payout.
- Trust: Gemini screening on submitted brand content.
- Team: Three full-stack engineers shipping end to end on AWS EC2.
Quantified Business Outcomes
Boksi unifies creator monetization and brand campaign operations with discovery, workflow, payments, and AI quality controls in one marketplace.
| Performance Metric | Before Boksi | Boksi Platform (Parallel Loop) |
| Discovery | Ad hoc outreach | Profiles + Demographic/Niche Filters |
| Order Workflow | Email and spreadsheets | Creator & Brand Dashboards |
| Payments | Off-platform transfers | Platform-Managed Payouts |
| Content Quality | Manual review only | Gemini AI-Generated Content Detection |
Technical Deep-Dive: Frequently Asked Questions
How do brand filters combine age, country, gender, and niche?
PostgreSQL stores structured creator attributes and niche taxonomies. Brand search queries apply composable filters in Express APIs, returning paginated profiles with engagement summaries. Saved filter presets let marketing teams rerun campaigns in the same segment without rebuilding queries.
How does the creator order workflow run from hire to payout?
A hire creates an order record with brief, deadline, and price. Creators upload deliverables to the order, triggering Gemini analysis. Brands approve or request revisions; upon approval, payment state moves to released according to platform rules defined in PostgreSQL and executed through the payment service on AWS EC2.
When does Gemini block or flag submitted brand content?
Each submission runs through Gemini API scoring for synthetic generation signals. High-risk deliverables are flagged in the brand dashboard with explainable scores so reviewers can reject before funds release. Low-risk authentic uploads clear faster, keeping campaign throughput high without sacrificing trust.
READY TO SHIP?
BOOK A 30-MINUTE CALL.
We'll discuss your idea, share a fixed-price quote, and map out a timeline. No sales pitch. No BS.