RAG v2

Name: RAG v2
Availability: InStock
Author: acidicsoil

Supports UI

by acidicsoil

Advanced document-grounded RAG for LM Studio with hybrid retrieval and answerability gating.

0 stars

0 views

Works in:claude

Exposes:Tools

View on GitHub Docs

What it does

Provides a sophisticated Retrieval-Augmented Generation (RAG) pipeline that enhances local LLMs with document-grounding. It utilizes hybrid semantic and lexical retrieval, multi-query expansion, and reranking to ensure high-precision evidence injection.

Tools

rag_answer: Generates a grounded answer based on the provided document corpus.
rag_search: Performs a targeted search across the corpus to retrieve relevant evidence.
corpus_inspect: Inspects the current state and metadata of the document corpus.
rerank_only: Specifically reranks a set of candidates to improve evidence quality.

Installation

Add this to your claude_desktop_config.json (adjust cwd to your local path):

{
  "mcpServers": {
    "rag-v2-local": {
      "command": "npm",
      "args": ["run", "mcp:stdio"],
      "cwd": "/path/to/rag-v2"
    }
  }
}

Supported hosts

Claude Desktop
LM Studio

Quick install

npm run mcp:stdio

Information

Pricing: free
Published: 5/12/2026
stars

Related Apps

Choose your AI client and follow the steps below.

Claude Desktop

{
  "mcpServers": {
    "rag-v2-local": {
      "command": "npm",
      "args": ["run", "mcp:stdio"],
      "cwd": "/path/to/rag-v2"
    }
  }
}