ArtisaNet

Live in Old and Enjoy now

Comment

Inspiration

My journey with ArtisaNet began in a bustling local market in my hometown. I watched artisans—potters, weavers, jewelers—pour their souls into handmade crafts, yet struggle to sell beyond a few passersby. One weaver told me:

I can make beautiful scarves, but I don’t know how to reach people online.

That hit me hard. In a world of Instagram influencers and e-commerce giants, these creators, who preserve cultural heritage, were invisible. I envisioned a platform like Instagram, but for products—where artisans could shine without needing tech skills. The idea was simple: use AI to automate marketing, letting artisans focus on creating.

What it Does

ArtisaNet is a mobile-first marketplace where local artisans upload photos and descriptions of their handmade products, and our AI auto-generates engaging marketing videos in their native language. These videos are posted to a social feed, like Instagram Reels, where buyers browse, bid, and buy. Key features include:

AI Video Creation: Artisans upload products; AI crafts 15–30 second videos with visuals, narration, and music, editable via simple prompts (e.g., "Add a festive vibe in Spanish"). Dynamic Bidding: Buyers bid to increase a product’s price, with the highest bidder getting first dibs. If unpaid within \( t_{\text{timeout}} = 24 \text{ hours}\), the next bidder steps in, ensuring fair value. Targeted Ads: AI tailors ads for specific audiences (e.g., eco-focused for green buyers, luxury for premium shoppers). Trend Alerts: Artisans get real-time insights (e.g., "Eco-crafts up 20%!") to optimize listings. Payments: Secure online payments via Stripe, with weekly payouts to artisans minus a 10–15% commission. ArtisaNet makes selling effortless, boosts visibility, and helps artisans earn more while preserving cultural crafts.

How We Built It

Frontend: React Native for a cross-platform app with an Instagram-like feed, ensuring accessibility for low-tech users. Backend: Firebase for user data, product listings, and payment integration with Stripe. AI Pipeline: Integrated Google Cloud Video AI for video generation, Stable Diffusion for visuals, and ElevenLabs for multilingual voiceovers. A fine-tuned mT5 model handles native-language prompts.

Challenges We Ran Into

AI Output Quality: Early AI videos lacked cultural nuance (e.g., generic visuals for traditional crafts). We fine-tuned models with artisan-specific datasets, but it took days of iteration. User Trust: Artisans were wary of tech platforms due to past exploitation. We demoed at a craft fair, showing real-time video generation to build confidence. Bidding Complexity: Initial auctions confused users. We added a "buy now" option and a minimum bid formula:

$$ P_{\text{min}} = C_{\text{cost}} + 0.1 \, C_{\text{cost}} $$

Connectivity: Many artisans had poor internet. We optimized for low-data usage and added offline upload queuing. Ethics: Ensuring AI didn’t mimic artisan designs was critical. Implemented watermarks and a "handmade-only" policy.

Accomplishments That We're Proud Of

Ethical AI: Designed an AI system that amplifies human creativity, not replaces it, setting a precedent for responsible tech. Built a Working MVP Flow: upload → auto-generate video → publish → open pitched-price window.

What We Learned

Empathy Drives Design: Artisans’ feedback shaped a simpler, voice-driven UI. AI’s Potential and Limits: Fine-tuning AI for cultural sensitivity is complex but critical for impact. Collaboration is Key: Pairing with artisans and open-source communities accelerated development and grounded our mission. Business Viability: A commission-based model balances affordability for artisans with platform sustainability. Resilience: Each challenge—whether tech glitches or user skepticism—taught us to iterate fast and stay mission-focused.

What's Next for ArtisaNet

Scale the Pilot: Partner with nonprofits to onboard 100 artisans in three months, targeting a 30% income uplift. Run a Pilot in One City: Onboard 50–200 artisans via NGO/co-op partnerships and measure time-to-first-video, conversion uplift, and payout satisfaction. Funding and Mentorship: Seek Hack2Impact mentorship and \$10,000 in seed funding to build a full-stack platform. Impact Metrics: Measure jobs created and traditions preserved, aiming to empower 1 million artisans by 2030. Support Multi-Languages: English, Hindi, Tamil, and more, with voice-to-text for low-literacy users.

Share this project:

Updates