llm-app

Why we included this project

Document Q&A and enterprise search only stay useful when they reflect the current state of your data, and this project skips the assembly work by shipping ready-to-run templates. The apps connect to sources like SharePoint, Google Drive, S3, Kafka, and PostgreSQL, and the indexes stay in sync as files change, so a deleted document stops showing up in answers. Indexing runs in memory and covers vector, hybrid, and full-text search, and the templates are built to scale to millions of pages. You can test them locally with Docker, then deploy to a major cloud or on-premises; switching a vector index to a hybrid one is a one-line change. For teams that want a working RAG pipeline without building the plumbing first, this is a practical place to start.