"Best open-source NPX or UVX tools for terminal deployment on Windows" matching MCP servers:

MCP server for compressing AI embeddings by 5-7x using TurboQuant (PolarQuant + QJL), with tools to compress, decompress, estimate savings, and embed+compress vectors.

Stores what AI agents learn — thoughts, decisions, session events and repo facts — in PostgreSQL with pgvector and serves it back through versioned MCP tools with hybrid vector-plus-full-text retrieval and budgeted context packs. Enables multiple agents to share one scoped, role-isolated long-term memory, with deliberate promotion of distilled candidates into durable memories.

Open-source MCP server for mem0 that enables memory storage and retrieval using local LLMs, self-hosted, and Docker-free.

Enables semantic search across Apple Mail, Messages, Calendar, and Contacts on macOS using natural language queries. All processing happens locally with privacy-first vector indexing for fast similarity search.

A self-hosted second brain MCP server that enables capturing thoughts with deduplication and semantic search using local embeddings and PostgreSQL with pgvector, all running on your own hardware.

A remote MCP server enabling semantic memory across multiple AI platforms like Perplexity Comet and Claude Desktop, with vector embeddings for intelligent retrieval.

Connects AI agents to Teradata databases for querying, analysis, administration, and AI pipelines with enterprise security.

Enables managing knowledge bases in Open WebUI and retrieving relevant fragments from Qdrant-stored knowledge through an MCP client.

A Machine Control Protocol (MCP) server that enables storing and retrieving information from a Qdrant vector database with semantic search capabilities.

An integration server implementing the Model Context Protocol that enables LLM applications to interact with Milvus vector database functionality, allowing vector search, collection management, and data operations through natural language.

Munin is a high-performance, pragmatic memory layer for AI agents (Cursor, Claude Code, OpenClaw, Gemini CLI,...). Unlike other solutions, Munin focuses on developer productivity with: * Multi-Project Support: Isolate memories into separate "brains" (Context Cores). * GraphRAG: Automatically builds a knowledge graph from your context. * Sub-200ms Search: Blazing fast Hybrid & Semantic

A local Retrieval-Augmented Generation system that enables users to ingest markdown files into a FAISS-powered vector knowledge base for semantic search. It provides tools for document indexing and context retrieval to support informed LLM queries without external dependencies.

Enables Claude to interact with core AWS services like S3, EC2, RDS, and CloudWatch, along with a generic SDK wrapper for any AWS operation. It also supports cost monitoring and optional vector store capabilities for document ingestion and search.

Enables LLM apps and AI agents such as Claude Desktop, Cursor, and Cline to build and query a knowledge base through standardized MCP tools, covering document chunking, vector embeddings, hybrid retrieval, and context injection. It can be launched instantly via uvx, Docker, or a local source install.

Enables natural language search and analysis of uploaded PDF, CSV, and Excel documents using retrieval-augmented generation and MCP tools, providing contextual answers to user queries.

Provides semantic search and complete document retrieval by ID through OpenAI's Vector Store API, letting deep research agents access company-specific knowledge and tools.

Enables users to ingest documents into a PostgreSQL/pgvector knowledge base, run semantic search over them, and get grounded answers through a retrieval-augmented generation pipeline backed by a free LLM. It also lets clients spin up specialized AI agents on demand and exposes knowledge-base stats and configuration as resources for tutoring workflows.

A local-first Graph-RAG system combining ChromaDB with metadata-based graph relationships and Gemini 2.5 Flash for intelligent Q&A over Obsidian vaults, supporting MCP clients like Claude Desktop, Cursor, and Raycast.

A server implementation that allows secure communication between MCP clients and privateGPT, enabling users to chat with privateGPT using knowledge bases and manage sources, groups, and users through a standardized Model Context Protocol.

Enables semantic search and document management with support for text, PDF, and image uploads using your own Supabase database and OpenAI API keys. Supports multi-tenant deployment on Cloudflare Workers or local hosting.

Agent memory on a local SQLite file or on PostgreSQL with pgvector, with provenance, trust verdicts and explicit abstention when memory cannot support an answer.

Enables AI agents to interact with an embedded graph database (GrafeoDB) via the Model Context Protocol, providing tools for graph CRUD, GQL queries, full-text and vector search, and graph algorithms.

Semantic memory for AI builders: capture the tacit engineering know-how that never reaches your docs, recall it the moment it applies. Built in Rust on Postgres and pgvector.

Official Telys MCP server: private, on-device AI memory and retrieval powered by native Mojo kernels. 19 tools: memory CRUD, semantic + BM25 search, single-key filters, compaction/IVF/tuning and self-refreshing repo auto-indexing. Stdio. Introspection needs no credentials; execution requires a one-time free telys login. Source: packages/telys-sdk/telys/mcp.py. Registry: io.github.thyn-ai/telys.

MCP server for OpenAI Vector Store API, managing vector stores, files, file batches, and semantic search.

Enables LLM hosts to retrieve live, relevant documentation excerpts from official library docs sites via a search-and-RAG tool, avoiding reliance on training data.

Provides tools for retrieving and processing documentation through vector search, enabling AI assistants to augment their responses with relevant documentation context.

Gives AI agents persistent, local-first memory using SQLite and on-device embeddings, enabling semantic search and recall across sessions with no cloud calls.

Provides SQL-backed semantic search over indexed notes using pgvector, exposing tools to search and list note sources via natural language.

An MCP server that enables LLMs to perform semantic and fulltext searches within Neo4j while executing complex, search-augmented Cypher queries for GraphRAG applications. It provides tools for database schema discovery and supports multi-provider embeddings to facilitate advanced graph traversals.

Local-first, source-traceable memory for AI agents — no LLM at ingest, $0 per message, zero data egress. Gives Claude Code, Cursor, and any MCP client one shared persistent memory with semantic recall, belief revision, selective forgetting, and a provenance guard that blocks acting on stale or unconfirmed memories.

Privacy-first local document search using semantic search. Runs entirely on your machine with no cloud services, supporting PDF, DOCX, TXT, and Markdown files.

Indexes local Markdown/text files into a SQLite database with vector embeddings and provides MCP tools for semantic search without cloud dependencies.

Agent memory for facts that change. When a user corrects something, Recall replaces the old fact, tells the agent what it replaced, and keeps the old value in a history. Ask what was believed on any date, see every change to a fact, or diff two dates. Runs locally on polign_db, Postgres with pgvector, MongoDB, or Qdrant. Apache 2.0.

Enables natural-language semantic search of a project's code through a local vector index, so questions that plain grep cannot match return ranked file:line references. It also indexes or refreshes projects on demand, reports index status, and optionally flags near-identical code across files.

An MCP server that provides AI assistants with access to Multi Theft Auto: San Andreas function documentation through vector similarity search and smart keyword expansion. It enables efficient information retrieval with features like deprecation warnings and SQLite caching for technical documentation.

Provides AI agents with persistent, searchable memory using semantic search, auto-linking, and categorization, with zero-config local setup or production-ready external providers.

Enables semantic retrieval over locally indexed PDF manuals and books, exposing tools to list available sources and search excerpts by query with page, chapter, source, and relevance metadata.

A local-first knowledge base for LLM coding agents that indexes repository documentation, concept ontology, and build targets into Qdrant and exposes retrieval as MCP tools (search, get, list sources, reindex).

Enables agents to run semantic search across one or more local project directories by automatically maintaining a LAN-local Qdrant index with Ollama embeddings. Indexing, staleness checks, and incremental updates happen transparently, so users can query code by meaning without managing collections, chunks, or hashes.