Tutorial: building a source-code RAG index with Spring AI and PgVector

On November 10, 2026, a tutorial published on DEV Community showed how to build a RAG index over a project's dependency source code using Spring AI and PgVector. The workflow deduplicates files, splits them into embedding-sized chunks, and stores them in a Spring AI vector store backed by PgVector. At query time, answers are grounded in the retrieved content and carry file citations taken from the retrieval metadata, so every response points back to the specific source file it came from.

The author frames the setup as a local, Context7-style index for the library sources a project actually uses, with the indexing process exposed over MCP so the index can be created and queried through that interface. The motivation given is timeliness: a model trained last year does not know about an API you upgraded to last week, and an index built from the current dependency sources closes that gap. The tutorial text available at publication time was cut off mid-description, so the full walkthrough was not yet complete at that point. As of now, this remains the only report on the topic, and there has been no follow-up coverage.