FieldCure.Mcp.Rag 1.3.0

FieldCure.Mcp.Rag

MCP RAG server with hybrid BM25 + vector search and AI-powered chunk contextualization — indexes documents from configured source paths, enriches chunks with AI-generated context and keywords, generates embeddings via OpenAI-compatible APIs, and performs keyword (FTS5) and semantic (cosine similarity) search with Reciprocal Rank Fusion.

Install

dotnet tool install -g FieldCure.Mcp.Rag

Architecture

fieldcure-mcp-rag
├── exec  --path <kb-path> [--force]   # Headless indexing (single KB)
└── serve --base-path <path>           # Multi-KB MCP search server (stdio)

• exec — headless indexing process. Scans source paths, chunks, contextualizes, embeds, stores.
• serve — read-only MCP server serving all KBs under the base path. Tools accept kb_id parameter to target a specific KB. Lazy-loads KB instances on first access.

Both modes read config.json from the knowledge base folder and resolve API keys from Windows PasswordVault.

Quick Start

1. Create a knowledge base

Create %LOCALAPPDATA%\FieldCure\Mcp.Rag\{kb-id}\config.json:

{
  "id": "my-kb-001",
  "name": "Project Docs",
  "created": "2026-04-03T00:00:00Z",
  "sourcePaths": ["C:\\Users\\me\\Documents\\project-docs"],
  "contextualizer": {
    "provider": "anthropic",
    "model": "claude-haiku-4-5-20251001",
    "apiKeyPreset": "Claude"
  },
  "embedding": {
    "provider": "openai",
    "model": "text-embedding-3-small",
    "apiKeyPreset": "OpenAI"
  }
}

2. Index

fieldcure-mcp-rag exec --path "%LOCALAPPDATA%\FieldCure\Mcp.Rag\my-kb-001"

3. Search (MCP server)

fieldcure-mcp-rag serve --base-path "%LOCALAPPDATA%\FieldCure\Mcp.Rag"

A single serve process handles all knowledge bases under the base path.

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "rag": {
      "command": "fieldcure-mcp-rag",
      "args": ["serve", "--base-path", "C:\\Users\\me\\AppData\\Local\\FieldCure\\Mcp.Rag"]
    }
  }
}

Tools (5)

All tools (except list_knowledge_bases) require a kb_id parameter.

Search Modes

Embedding is optional. Without it, BM25 keyword search still works.

Supported Formats

DOCX, HWPX, XLSX, PPTX, PDF (with OCR fallback for scanned pages), TXT, MD — auto-extends when new parsers are added to FieldCure.DocumentParsers.

config.json Reference

Data Storage

Knowledge base data at %LOCALAPPDATA%\FieldCure\Mcp.Rag\{kb-id}\:

• config.json — knowledge base configuration (created by app)
• rag.db — SQLite database (chunks, embeddings, FTS5 index, file hashes, indexing lock)

Requirements

• .NET 8.0 Runtime or later

See Also

Part of the AssistStudio ecosystem.

Links

• GitHub
• MCP Specification