Self-hosted alternatives to rag-architect

R

rag-architect

Designs and implements production-grade RAG systems by chunking documents

rag-architect is an agent skill from the jeffallan claude-skills collection for designing and building retrieval-augmented generation systems, which answer questions from your own documents. The skill covers the whole pipeline: chunking documents, generating embeddings, configuring vector stores, building hybrid search and reranking, and measuring retrieval quality. Its workflow has five stages. The agent first analyses the retrieval needs, latency limits, accuracy requirements and scale. It then designs the vector store, choosing the database, the schema, the indexing strategy and the sharding approach. The chunking stage decides how documents are split, how much they overlap, where semantic boundaries fall and what metadata to attach. The retrieval pipeline selects an embedding model, transforms queries, combines keyword and vector search and reranks results. The final stage tracks metrics, debugs poor retrieval and keeps optimising. The skill says to validate at each step before moving on, and it points to reference material for each choice. It suits teams building search or question answering over internal documents who want a retrieval setup designed deliberately. It is listed on skills.sh, where the page shows install counts and GitHub stars for the collection. Its Gen Agent Trust Hub audit is marked as failing, while Socket and Snyk show warnings, which the FAQ lists. Install it with npx skills add https://github.com/jeffallan/claude-skills --skill rag-architect.