MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
mcp-rerankergithub.com/globalpocket/mcp-rerankerupdated this year
A generic Model Context Protocol (MCP) server for high-accuracy document reranking using sentence-transformers (Cross-Encoder). Ideal for enhancing RAG and AI agent decision-making. sentence-transformers (Cross-Encoder) を使用して文書の関連度を再計算・ソートする汎用 MCP (Model Context Protocol) サーバー。
Apache-2.0 ★ 1 -
Cohere-compatible reranker over MCP — pay-per-call in USDC, no account, no GPU, no model download.
pypi MIT -
Self-hosted Qwen3 retrieval server with Embedding, Reranker, MCP, and Web console.
★ 7 -
Local-first RAG engine with hybrid search, BGE reranker, self-learning loop, and MCP server
-
One MCP server for your agent's memory: consolidate a markdown knowledge base's tools behind one surface, and rerank retrieval.
MIT -
MCP server for DeepSeek native web research, multi-provider web search, SearXNG fallback, and optional OpenRouter reranking.
MIT ★ 2 -
Local help-desk RAG: hybrid retrieval, reranking, evals, tracing and an MCP tool, all on a laptop (Ollama, Qdrant, Phoenix)
MIT -
MCP search server for a self-hosted SearXNG, with BM25F ranking and optional decision-model reranking
MIT -
poisonvectorsgithub.com/jamesburton/poison-vectorsupdated this year
MCP server for semantic vector search with relevance feedback (poison-vector technique). Exposes embed, rerank, and feedback tools over stdio.
nuget -
Local hybrid-search (BM25 + vector + reranking) MCP server — ask questions of any local codebase or document folder, fully offline
pypi MIT ★ 9 -
scientific-rag-mcpgithub.com/mikrominiw/scientific-rag-mcpupdated this year
Bilingual (RU/EN) scientific-literature RAG: vision PDF parsing, BGE-M3 hybrid retrieval, cross-encoder reranking, over MCP.
MIT -
oboyugithub.com/sonesuke/oboyuupdated this year
Self-hosted MCP Japanese text indexing & search—chunking+embeddings with BM25×vector rerank
pypi MIT ★ 3 -
Folio: agentic RAG MCP server with hybrid search, rerank and contextual retrieval on pgvector, behind MCP Gate tokens
MIT -
memorymcpgithub.com/gashel01/memorymcpupdated this year
Persistent, tiered cognitive memory for AI agents, with optional cross-encoder reranking. Python lib · CLI · MCP server · FastAPI.
pypi Apache-2.0 ★ 1 -
Servidor MCP para vault Obsidian: busca híbrida local com reranker, escrita restrita ao Inbox e evals de retrieval, tool use e injeção de prompt
MIT -
Pluggable, observable RAG MCP server — hybrid retrieval (BM25+dense+RRF), BEIR-benchmarked rerankers (BGE / TypeSafe Jev / LLM), Streamlit dashboard + Langfuse, multimodal ingestion
-
Index local files for semantic search and reranking using an MCP server, CLI, or web console on your machine.
MIT -
Multi-engine web search + fetch MCP server (DDG, Brave, Tavily, Google AI Mode) with consensus-merge dedup, local cross-encoder reranking, and Groq synthesis
pypi MIT ★ 1 -
Searchable memory for your Claude Code sessions: local hybrid search with reranking, a CLI, and an MCP server for agents.
npm MIT -
langsearch-mcp-servergithub.com/fusman60/langsearch-mcp-serverupdated this year
MCP server for LangSearch Web Search & Semantic Rerank API. Use with Claude Code, Claude Desktop, and any MCP-compatible client.
npm MIT -
Local-first, Obsidian-aware RAG MCP server — hybrid retrieval (dense + BM25 + RRF), contextual retrieval, and cross-encoder reranking on a Clean Architecture skeleton.
MIT -
Local OpenAI-compatible embeddings and reranking on Apple Silicon: bge-m3 hybrid search, Qwen3, a retrieval eval harness and MCP memory for agents.
MIT ★ 1 -
An MCP server that gives any LLM host grounded, cited retrieval over your own documents — hybrid BM25 + dense retrieval, cross-encoder-ready reranking, and built-in eval.
npm MIT -
zemailgithub.com/zeroentropy-ai/zemailupdated this year
Semantic search over your Gmail, from inside Claude Code. Embeds your full inbox with zembed-1 and reranks with zerank-2 — finds emails by meaning, not keywords.
★ 5 -
MCP server exposing NaN API media tools: image generation/editing (flux-2-klein), TTS (kokoro), STT (whisper), embeddings and rerank
npm MIT ★ 12 -
Evidence selection and quota-aware recovery for Codex. Optional Jev reranking, local privacy controls, exact source reads. Keep your coding model.
MIT ★ 2 -
Private web search MCP: SearXNG metasearch → ML reranking → 4-tier fetch cascade (Firecrawl / Crawl4AI / raw / Wayback) with per-domain learning, Ollama query expansion and synthesis. No cloud APIs.
npm MIT ★ 29 -
ragPapergithub.com/i1ight/ragPaperupdated this year
Local-first paper RAG + read-only MCP server: hybrid/reranker retrieval, self-correcting DOI enrichment, and citation graphs over your Zotero/PDF library — queryable from Claude Code, Codex, or any LLM.
MIT ★ 2 -
yinjugithub.com/Gor-c/yinjuactive
A knowledge base that would rather say 'I don't know' than make something up. Hybrid retrieval, rerank, refusal gates, MCP. Go + Vue 3.
MIT -
Production embedding & decision inference service in Rust — Candle/ONNX dual engine, CUDA/Metal acceleration, REST/gRPC/MCP/CLI generated via sdforge, rerank, Matryoshka dimensions, OpenAI-compatible API, and semantic caching
Apache-2.0 ★ 1 -
Learn how Retrieval-Augmented Generation actually works — then run one. An interactive, animated RAG lab: hybrid retrieval (dense + BM25 + RRF + rerank), a live web UI, and an MCP server.
MIT -
A retrieval engine that refuses to guess. Register git repos and URLs, get grounded, cited answers over MCP and REST. Hybrid search, cross-encoder rerank, Postgres + pgvector.
MIT -
Intent-first search that finds the right tool from a vague description. Hybrid retrieval over LLM-written cards, LLM rerank with recall safety, MCP server for AI agents, runs on a laptop CPU.
MIT ★ 1 -
Literature knowledge base for DSH and MCP clients. Index PDFs or a Zotero library into one SQLite file; search and RAG tools return passages with section, page and DOI. Embedding and reranking run locally.
npm MIT -
Code, documentation, and public web retrieval for SDK and MCP clients, with optional embeddings and reranking. Returns source-backed context; the calling agent stays in control.
npm MIT ★ 1 -
Local-first, deterministic memory store for AI agents: SQLite-backed BM25 with optional local ONNX semantic search and cross-encoder rerank, graph projection, and an MCP adapter. No required network or LLM.
cargo Apache-2.0 ★ 4 -
Local code search combining BM25, vector similarity, and cross-encoder reranking. Parses 60+ languages with tree-sitter, runs entirely offline, and returns structured results with file paths, line ranges, and symbol metadata. Built in Rust.
pypi MIT ★ 150 -
codeseekupdated this year
Code intelligence CLI for Claude Code — AST call graph analysis and hybrid semantic search (Dense + Sparse + RRF + Cross-Encoder Reranker) across 7 languages. Ships as MCP tools for Claude Code & Codex CLI.
npm MIT -
Enterprise Multi-Agent Decision Intelligence Platform powered by LangGraph, Clause-Aware Hybrid RAG (Milvus + BM25 + BGE Reranker), Model Context Protocol (MCP) NL2SQL with SQLGlot AST guardrails, and FastAPI.
Apache-2.0 -
Read-only MCP server for 1C BSL code search: hybrid BM25 + vector kNN retrieval over methods in Elasticsearch, RRF, dedup, rerank — context for code LLMs. | MCP-сервер поиска по коду 1С BSL: гибридный BM25 + векторный поиск по методам в Elasticsearch, RRF, дедупликация, ранжирование.
★ 3 -
Self-hosted MCP server for web search using SearXNG (40+ engines) with CrossEncoder reranking // MCP server de búsqueda web local con SearXNG y reranking semántico (CrossEncoder + BM25 + RRF)
-
web-mcpgithub.com/ixchio/web-mcpupdated this year
Web MCP Server + RAG Pipeline with Cross-Encoder Reranker, FAISS, and Local LLM
MIT -
Academic-paper knowledge-base MCP server: PDF ingestion (MinerU), Qdrant vector search with reranking, two-phase KB management
pypi MIT ★ 5 -
Local-first memory MCP server with local embeddings, vector search, and reranking
npm MIT -
mcp-rag-pgvector-demogithub.com/fr3kchy/mcp-rag-pgvector-demoupdated this year
Production RAG pipeline: pgvector + nomic-embed-text + Claude API with MMR reranking and eval harness
★ 1 -
task-fabricgithub.com/dlhck/task-fabricupdated this year
File-based, git-versioned task manager for AI agents — plain markdown tasks, hybrid search (BM25 + semantic + LLM reranking), MCP server over Streamable HTTP
-
Unified MCP search server with 8 providers, caching, reranking, and rate limits
npm MIT ★ 20 -
greb-jsupdated this year
JavaScript port of op-grep - intelligent code search with LLM reranking
npm MIT -
Retrieval for Markdown vaults: hybrid keyword and semantic search with reranking, as a CLI, terminal UI and MCP server.
MIT -
A Model Context Protocol (MCP) server providing AI-powered Baidu search with intelligent reranking and web content extraction
pypi MIT License -
Local semantic search over your own files, with an MCP interface for coding agents. pgvector recall, LLM rerank for precision.
MIT -
Local RAG workspace with persistent documents, hybrid search, reranking, cited answers, FastAPI, MCP and LangChain.
-
tokengramgithub.com/kinleee/tokengramupdated this year
Token-saving memory layer for AI coding assistants — local hybrid RAG (BM25 + semantic + reranker) over your project, exposed via MCP.
pypi LicenseRef-Proprietary-Tokengram -
Config-schema-driven vector knowledge base: LanceDB storage, pluggable embedding/reranker/graph backends, CLI + HTTP REST + MCP (stdio & streamable HTTP)
cargo MIT OR Apache-2.0 -
books_raggithub.com/zcoderz/books_ragupdated this year
Precision-first RAG engine for technical books — hybrid retrieval, reranking, citation-backed answers with abstention, 34 MCP tools
Apache-2.0 ★ 1 -
RAG-System-from-Scratch-avec-MCPgithub.com/FEDI-HASSINE/RAG-System-from-Scratch-avec-MCPupdated this year
Système RAG complet construit de zéro avec recherche sémantique, reranking, agent IA, API MCP, évaluation LLM-as-a-Judge et interfaces CLI/Streamlit.
★ 1 -
RAG server for quantitative finance research — layout-aware PDF/LaTeX/notebook ingestion, hybrid BM25 + BGE-M3 retrieval with reranking, local LLM synthesis, REST API and MCP server.
★ 2 -
mcp-servergithub.com/gpu-bridge/mcp-serverupdated this year
GPU-Bridge MCP Server — 30 AI services as Model Context Protocol tools. LLMs, image gen, video, audio, embeddings, reranking, PDF parsing & more. x402 native for autonomous agents.
npm MIT ★ 2 -
cheetah-grebupdated this year
AI-powered intelligent code search with LLM reranking - MCP server for Claude, Cursor, Cline
npm MIT -
cheetah-grebupdated this year
AI-powered intelligent code search - client package with grep, read, and AST analysis (GPU reranking via cloud API)
pypi MIT