Lanjut ke konten
image-caption

Sistem Knowledge Retrieval RAG

Pipeline RAG menggunakan Solr dan Vector DB untuk konteks LLM yang akurat.

01 Demonstration / Case Study

*Automated Alt-Text: Every new image was automatically processed to generate descriptive, SEO-friendly captions.

*Smart Tagging: The AI categorized items not just by "Shirt," but by "Breathable linen summer shirt, slim fit, navy blue," automatically filling in database metadata.

*Visual Search: Users could upload a photo of a style they liked, and Gemini would "reason" through the visual features to find matching categories in the shop.

*Accessibility: 100% of the site became WCAG 2.1 compliant within weeks.

*Efficiency: Manual data entry time was reduced by 92%.

*SEO Boost: More descriptive, keyword-rich captions led to a 15% increase in organic search traffic.

Model Performance

model gemini-2.5-flash
vector db Pinecone
latency 800ms

Tech Stack

GeminiAI Nextjs Supabase MongoDB

Access

Instant temporary access for demo.