Launch HN: Captain (YC W26) launches automated RAG for files, aiming to simplify unstructured data search. The tool indexes cloud storage like S3 and GCS, plus SaaS sources like Google Drive, handling everything from text extraction to semantic search.

Building production RAG pipelines manually requires substantial effort - teams must manage ETL, chunking, embedding, storage, search, re-ranking, and compliance while optimizing for latency. Many DIY pipelines produce inconsistent results even after weeks of tuning. Captain automates this entire process through a single API call that indexes URLs, cloud storage buckets, directories, or individual files.

The system uses advanced techniques including Gemini 3 Pro for images, Reducto for complex documents, and Extend for OCR. After switching from 'gemini-embedding-001' to Contextualized Embeddings from 'voyage-context-3', the team achieved better results by encoding chunk embeddings with surrounding document context awareness.