Engineering notes — things I keep re-explaining, written down once.
- Chatbot retries, rate limits, and billing: one gateway or direct vendor accounts?
- Cost-sizing a moderation queue: token math for user text and images in Node.js
- RAG summarization in Node.js: PDF pages, embeddings, rerank, then a final summary
- Ask Your Docs: Node.js Semantic Search, Embeddings, Rerank, and RAG
- OpenAI-compatible image generation in Node: multi-provider fallback and model routing
- Picking an API to summarize multilingual support tickets, emails, and meeting notes
- One API key for OpenAI, Claude, and Gemini: text classification routing in Node.js
- Low-cost AI chatbot backend for a SaaS startup: token pricing, batching, caching
- Hybrid search for a docs chatbot: keyword plus embeddings, then rerank in Node.js
- Why your ask-your-docs RAG chatbot answers wrong despite good embeddings