Rag markdown retrieval
AS
2d ago+167+167+012345678901234567890123456789-120-120-012345678901234567890123456789
PR Checklist
Please check if your PR fulfills the following requirements:
- The commit message follows our guidelines: https://voltagent.dev/docs/community/contributing/#commit-convention
Bugs / Features
- Related issue(s) linked
- Tests for the changes have been added
- Docs have been added / updated
- Changesets have been added https://voltagent.dev/docs/community/contributing/#creating-a-changeset
What is the current behavior?
What is the new behavior?
fixes (issue)
Notes for reviewers
Summary by cubic
Rewrites the RAG chatbot example to retrieve answers from local Markdown documents instead of a hardcoded knowledge base, and aligns the quick-start docs with what the app generator actually produces.
- The retriever now scans
documents/*.md, splits files by Markdown heading, ranks sections by keyword overlap, and returns the top three matches with their source file and heading. - Matched
{ source, section }entries are stored in the operation context underreferencesand included in the text sent to the model. - Adds sample documents and unit tests for the Markdown search, and updates the example README so citations point to the new files.
Summary by CodeRabbit
- New Features
- The RAG chatbot example now searches local Markdown documents and uses relevant sections to answer questions. Answers include references to their source documents, and the example reports when no matching content is found.
- Documentation
- Updated setup guidance to require Node.js 20 or newer and explain where to add provider API keys when skipping the setup prompt.
- Added details about the chatbot’s document search and references, plus introductory information about agents, tools, and workflows.