Ragdrift MCP runs on your own machine — the client starts it, so there is no endpoint to ping. 26 installs a week from npm. Last commit 13 Jun 2026.
Diagnose RAG drift: interpret scores, recommend thresholds, explain dimensions.
Today is the operative word: we check Ragdrift MCP every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add ragdrift-mcp -- npx -y @mukundakatta/ragdrift-mcp
{
"mcpServers": {
"ragdrift-mcp": {
"args": [
"-y",
"@mukundakatta/ragdrift-mcp"
],
"command": "npx"
}
}
}
[mcp_servers.ragdrift-mcp]
command = "npx"
args = ["-y", "@mukundakatta/ragdrift-mcp"]
{
"mcpServers": {
"ragdrift-mcp": {
"args": [
"-y",
"@mukundakatta/ragdrift-mcp"
],
"command": "npx"
}
}
}
{
"mcpServers": {
"ragdrift-mcp": {
"args": [
"-y",
"@mukundakatta/ragdrift-mcp"
],
"command": "npx"
}
}
}
Market regime, 7-dimension risk vectors and payment hazard grades. Abstains when data is thin.
Author rules from policy docs, then decide: a Rete engine gives the verdict, an LLM explains why.
Generate QA datasets & evaluate RAG systems with failure diagnosis. Any LLM.
Citation-guarded retrieval over 22M Taiwan court judgments and administrative interpretations
Production-ready RAG + MCP demo: eval-in-CI merge gate, Langfuse traces, structure-aware chunking.
MCP RAG: index PDFs, repos, YouTube, Discord, text; optional YouTube vision; query with citations.
PowerPoint (.pptx) MCP server: editable vector diagrams, tables, budgeted reads, render-to-verify
Ingest, manage, and retrieve documents for RAG-powered AI applications
Answers built from our own checks of this server.