$ loading_
Run hybrid search over self-hosted public data indexes via MCP.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "windex MCP server" yet — see the docs or source repo.
Use the windex MCP server to search indexed GitHub, devdocs, and small web data for materials about vector retrieval and hybrid search implementation, then summarize them by source.
Returns relevant results with a source-grouped summary for GitHub, docs, and web pages.
Search the continuously ingested news and Hacker News indexes for recent content related to open-source AI infrastructure, and list the most relevant topics and sources.
Provides relevant news and community discussion results with topic clustering and source information.
Use windex MCP to search the arXiv and Wikipedia indexes for information on combining sparse and dense retrieval, and output key concepts with references.
Returns relevant papers and encyclopedia entries, highlighting core concepts and references.
Developers can connect this MCP tool to an LLM workflow so the model queries self-hosted public-data indexes instead of relying only on general web search. It fits cases that need controlled data sources and a custom embedding model.
Researchers or analysts can search across indexed news, GitHub, Wikipedia, arXiv, developer docs, and Hacker News in one place. This helps them gather signals from multiple sources faster.
When public datasets are continuously ingested into the indexes, users can search newer content through this tool. It suits technical research and intelligence-gathering tasks that depend on changing information streams.
It provides hybrid retrieval over self-hosted indexes through the Model Context Protocol, combining dense and sparse search. The description says it can search continuously ingested public datasets such as news, GitHub, Wikipedia, arXiv, devdocs, and Hacker News.
The description indicates it uses “your own embedding model.” The provided material does not specify which models are supported or how to configure them; see the source repository.
It focuses on hybrid search over self-hosted indexes rather than directly calling a general-purpose search engine. Its scope is the ingested public datasets, and it is exposed to models through MCP.
Index local documents and run hybrid semantic-keyword search on-device.
Search documents with hybrid retrieval, reranking, guardrails, and tracing.
Enable self-hosted deep web research for AI agents without API keys.
Lets MCP clients search the web and fetch pages without API keys.
Search markdown knowledge bases with hybrid ranking and intelligent reranking.
Use MCP to query knowledge hubs, search collections, and summarize documents.