Anúncio da integração nativa de RAG e Web Search da OpenAI no n8n…

INEMA

NEW n8n adds OpenAI native RAG + Web Search (very helpful) n8n just integrated OpenAI's built-in tools directly into their platform - you can now do RAG and web search without connecting Pinecone, Supabase, or Perplexity CRITICAL: only works on the latest beta version (1.173+) - older versions will run endlessly and never complete, so update your workspace first Here's how RAG works: go to platform.openai.com

• → upload your files
• → get a vector store ID
• → paste that ID into n8n's file search tool
• → ask questions and get cited responses in 30-50 seconds

OpenAI automatically formats the responses with citations and indexes, so you skip the manual cleanup steps you'd normally need Web search is also built-in now with adjustable content size (low, medium, high) - takes 20-40 seconds per search and returns URLs, citations, and formatted answers The honest limitations: OpenAI's RAG is more surface level than custom Supabase or Pinecone setups with less control over chunking, and web search is slower than Perplexity's sonar models Requires GPT-4o or higher (GPT-4.1, GPT-5) to work properly

RAG e Pesquisa Nativa no N8N

1