Rag
Add hybrid-search RAG with multiple knowledge bases and AI chunk contextualization so your coding agent grounds answers in your docs and code notes.
Overview
io.github.fieldcure/rag is a MCP server for the Build phase that provides hybrid-search RAG, multi-KB support, and AI chunk contextualization for agents.
What is this MCP server?
- Hybrid search across corpora with multi-knowledge-base support
- AI-powered chunk contextualization for sharper retrieval snippets
- Optional provider keys: OpenAI, Anthropic, Gemini, and Voyage (per server schema)
- FieldCure.Mcp.Rag v2.5.1 NuGet package with stdio transport
- Suited for doc-heavy solo products and internal playbooks
- Server version 2.5.1
- Documented env keys include OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY, VOYAGE_API_KEY
- Transport: stdio via NuGet FieldCure.Mcp.Rag
What problem does it solve?
Your agent keeps hallucinating project details because your docs and specs are too large to paste into every chat.
Who is it for?
Solo builders running doc-heavy agents who want multi-KB RAG with hybrid retrieval and optional multi-vendor model keys.
Skip if: Simple one-repo tasks where basic file grep or a single static rules file is enough.
What do I get? / Deliverables
After install, the agent queries indexed knowledge bases with hybrid search and richer contextualized chunks.
- Hybrid-search MCP query tools over one or more KBs
- Contextualized retrieval snippets for agent prompts
Recommended MCP Servers
Journey fit
RAG MCP servers are wired during Build when you equip agents with retrieval over project knowledge, not during initial idea brainstorming alone. agent-tooling is the right shelf for retrieval, embedding, and multi-KB MCP capabilities that extend model context.
How it compares
Full RAG MCP stack with hybrid search, not a lightweight single-file context skill.
Common Questions / FAQ
Who is io.github.fieldcure/rag for?
Developers using MCP agents who need searchable knowledge bases with hybrid retrieval and contextualized chunks over their own content.
When should I use io.github.fieldcure/rag?
Use it during build and ongoing implementation when grounding codegen and answers in indexed docs outweighs raw prompt context.
How do I add io.github.fieldcure/rag to my agent?
Install FieldCure.Mcp.Rag via NuGet, configure stdio in your MCP client, define knowledge bases, and set OPENAI_API_KEY and/or other listed provider keys as needed.