Mcp Knowledge Rag
Wire permission-aware document RAG into your agent so answers cite ingested docs via hybrid search on a hosted API.
Overview
io.github.infoinlet-marketplace/mcp-knowledge-rag is a Build-phase MCP server that connects agents to permission-aware RAG with ingest, hybrid search, and cited answers via a hosted API.
What is this MCP server?
- Permission-aware retrieval so agents respect doc access boundaries
- Ingest documentation and query with hybrid search
- Cited answers designed for trustworthy agent replies
- Hosted API backend with stdio MCP front door @infoinlet/mcp-knowledge-rag v0.1.0
- Pairs with Claude Code or Cursor when product knowledge must stay current
- npm package @infoinlet/mcp-knowledge-rag version 0.1.0
- stdio MCP transport; server schema dated 2025-12-11
- Capabilities: ingest docs, hybrid search, cited answers (hosted API)
What problem does it solve?
Agents guess when your specs live in scattered docs and you cannot enforce who should see which knowledge base.
Who is it for?
Builders adding grounded Q&A to coding agents over proprietary docs, policies, or API references with attribution requirements.
Skip if: Teams that need fully offline RAG with no hosted API or who only need a single static file in context.
What do I get? / Deliverables
After registration, the agent retrieves from ingested corpora with citations and permission checks instead of improvising from stale prompts.
- Searchable ingested corpora exposed through MCP tools
- Hybrid retrieval results with citation-friendly agent responses
- Permission-scoped answers aligned to your doc ACLs
Recommended MCP Servers
Journey fit
Retrieval stacks are assembled while you shape agent workflows and knowledge boundaries—primary home is Build agent-tooling, not distribution or ops monitoring. Ingest, hybrid search, and cited answers are core RAG plumbing for coding agents, which maps directly to agent-tooling integrations.
How it compares
Permission-aware RAG MCP gateway, not a general web-search skill or raw vector-DB admin tool.
Common Questions / FAQ
Who is mcp-knowledge-rag for?
Indie developers and small teams building agent workflows that must answer from controlled document sets with citations and access rules.
When should I use mcp-knowledge-rag?
Use it in Build while wiring agent-tooling so product, support, and engineering docs stay searchable with hybrid retrieval before you ship customer-facing bots.
How do I add mcp-knowledge-rag to my agent?
Install @infoinlet/mcp-knowledge-rag, configure stdio in your MCP client, and complete hosted API credentials plus ingest for the collections you want searchable.