Serve production agentic RAG through R2R
Ingest documents, expose hybrid retrieval and agentic RAG APIs, and give agents a production retrieval service with citations and reviewable document operations.
npx skills add agentskillexchange/skills --skill serve-production-agentic-rag-through-r2r
Use R2R when an operator needs a production retrieval service behind an agent workflow, not just an embedding library or a one-off vector search script. The workflow is concrete: deploy the R2R server, ingest documents through the API, configure retrieval and generation behavior, run search, RAG, or agentic retrieval calls, and review answers with source-backed context. This is especially useful when multiple assistants need the same maintained knowledge layer with document operations, hybrid search, citations, and API access. The scope boundary is the operated retrieval service and API workflow: this is for teams that want agents to query a maintained RAG backend instead of wiring raw document parsing, vector storage, search, and generation directly into every assistant.
What this skill actually does
Inputs and prerequisites: Python or Docker, R2R server, API keys for the selected model provider, optional JavaScript SDK.
Setup notes: Install with `pip install r2r`, set the required model provider API key, run `python -m r2r.serve`, then connect agents or clients to the R2R API for document ingestion, search, RAG, and agentic retrieval calls.
Source and verification boundary: use https://r2r-docs.sciphi.ai/ as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.