The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
I built a working AI support inbox this afternoon, and I want to know what would make you trust it.
Relay reads every inbound support email, sets priority and category, writes a one-line summary and drafts the reply. A human approves with one click. Next.js, Prisma and Claude, with a rules engine underneath so the demo runs even without an API key.
Honest question for anyone who has worked a support queue: what would an AI draft need to show you before you'd hit send without reading the whole thread? A confidence score? The lines it based the answer on? Or nothing, you'd always read it?
Live demo goes up this week, case study after.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Challenges
View all
Related posts
More value from the energy you generate and store. Wattbank helps homeowners manage solar panels and home batteries using electricity prices and weather forecasts to plan when to charge, power the home, or sell surplus energy.
The landing page connects an immersive, nature-inspired visual direction with a clear product story, showing how smarter energy management works throughout the day and how operators can manage their communities.
View the project on Dribble:
A graphics card, drawn only with lines.
Hover to see it come apart layer-by-layer: backplate, PCB, heatsink, shroud, fans.
I turned it into a small playground.
Move your cursor across the layers, use the intensity slider to widen the gaps, and toggle light/dark.
Built with Claude Code and the Hairline skill by Lucas Marques, then ported to Next.js. Pure SVG, no 3D library.
Playground link in the comments 👇
I really love the illustration
Enterprise RAG AI: Private Doc Intelligence
A fast-growing B2B software firm needed a secure, enterprise-grade Retrieval-Augmented Generation (RAG) system to query thousands of confidential internal documents (legal contracts, technical architecture specs, Notion pages, and employee SOPs) in real time without third-party data leakage.
Key Challenges:
High hallucination risk and irrelevant context retrieval using standard keyword search.
Managing multi-format files (scanned PDFs, tables, markdown) with strict access control based on user hierarchy.
Technical Implementation & Solution:
• Engineered an automated ingestion pipeline in Python with LlamaIndex and LangChain for semantic chunking and metadata preservation.
• Integrated a hybrid vector search architecture pairing dense embeddings (OpenAI text-embedding-3) with sparse BM25 indexing in Qdrant.
• Added a cross-encoder re-ranking layer (Cohere Re-ranker) to prune low-relevance chunks, reducing API token costs by 40%.
• Built robust Role-Based Access Control (RBAC) via FastAPI endpoints, ensuring departmental data isolation.
Results & Impact:
• Dropped internal information retrieval time from 20 minutes to under 4 seconds.
• Attained 98% factual precision with traceable in-line page citations.
• Full on-prem Docker deployment ensuring 100% private data security.
I'd test revoking a user's document access mid-conversation. In your RBAC flow, do cached retrieval results get rechecked against current permissions before they're sent to the reranker or model?