Rapid Prototyping in Next.js: Building a Viral Ai Tool in 8 Hours
summary
Claims to have built Remembrall in 8 hours: a long‑term memory proxy that inserts remembered context into LLM calls with a one line change to OpenAI requests. Architecture runs on Vercel Edge Functions, computes embeddings, queries Postgres as a vector DB, duplicates streaming responses, and makes a lightweight model call to extract JSON memories; logs go to TinyBird. Frontend uses Next.js App Router, T3 stack (tRPC, Prisma, Tailwind) and ShadCN UI. Project is open‑sourced.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 0:50The Problem
- 1:50Demo
- 3:10How I built it
- 4:29Walkthrough
- 7:15Architecture
- 8:02Key takeaways
related talks
10mRapid Prototyping with AI
24mFull Stack Web Development with AI
48mGoing Beyond Prompt Engineering
46mBecoming an AI engineering company
42mPrompt. Build. Repeat. – The Vibe Coding Mindset
20mFully Integrated AI that Actually Ships
27mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
25mBuilding high performance, AI-native engineering teams