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.
-
Self-hosted web search and document understanding for AI agents
MIT ★ 2 -
Self-hosted Obsidian-style markdown vault with a web UI and a shared-memory MCP server for Claude Code
MIT -
Self-hostable web fork of Agentero: headless ACP agent host, MCP server and provider-abstracted RAG (WeKnora). AI-generated code; no maintenance guarantee.
MIT -
Self-hosted web search, page reading and field extraction for coding agents: an HTTP and MCP API over SearXNG with a shared cache and an egress proxy.
MIT -
web-searchgithub.com/idefav/web-searchupdated this year
Self-hosted browser-backed web_search and web_fetch for Codex, Claude Code, OpenCode, Pi, OpenClaw, HermesAgent, LangChain, and custom Agents — MCP + REST with provider fallback and SSRF protection.
MIT ★ 7 -
searxng-crawl4ai-mcpstale
Self-hosted MCP server for web search (SearXNG) and scraping (Crawl4AI)
npm MIT -
Self-hosted web search and page extraction for AI agents. No API key, no account, no telemetry.
Apache-2.0 -
Self-hosted web scraper for AI agents. Turn any URL into clean Markdown, crawl and map sites, search the web, extract JSON. Python, FastAPI, Playwright, MCP server and CLI.
MIT -
Anonymous-capable, self-hosted, account-free web search and fetch for agents. CLI + MCP (built on incur), pi-agnostic. Swappable backend and egress (direct, http proxy, socks5/Tor).
npm AGPL-3.0-or-later -
Self-hosted SrvDeck control plane: web console, multi-node management, authorization, audit, and MCP. In development.
AGPL-3.0 -
Self-hosted web analytics you query by asking an AI agent, instead of reading a dashboard.
MIT ★ 4 -
localbrowsergithub.com/iangogentic/localbrowserupdated this year
Self-hosted web search + page reading for local LLMs. SearXNG + fastCRW over MCP. No API keys, no per-query cost.
NOASSERTION -
松果 Acorn:完全属于你的信息系统——订阅、阅读、批注、翻译、本地存档。A self-hosted reader and personal web archive.
AGPL-3.0 -
Self-hosted web search and page extraction for AI agents. Stdio MCP server, HTTP API and Python/npm CLI.
pypi MIT ★ 2 -
Open-source, self-hosted web pentesting suite with agentic scanning, analyst tools, API testing, model checkpoints and MCP integrations.
Apache-2.0 ★ 386 -
Self-hosted, evidence-first web research stack for any AI agent — SearXNG, extraction, browser rendering, HTTP, Python SDK, and 16 MCP tools.
MIT -
Open-source, revenue-first web analytics on your own Cloudflare. Self-hosted DataFast alternative with MCP usage tracking. One-click deploy.
MIT -
Self-hosted web search MCP server (Node). 6 engines (CN-direct + Intl-via-proxy auto-routing), RRF fusion, SimHash dedup, per-host rate limiting, frequency-aware cache, TLS/JA3 impersonation fallback. Tools: web_search / fetch_url / deep_research.
MIT -
Web access over Tor for AI agents via MCP — a new circuit per call, no DNS leaks, .onion support. Self-hosted, MIT.
MIT ★ 2 -
Self-hosted Rust AI engine, token-saving tree-sitter coding CLI, and MCP autonomous agent server by Muhammad Afzal Kalwar (INDUS Web Agency)
MIT -
agent-searchgithub.com/cedarsaam/agent-searchupdated this year
Self-hosted, MCP-native web-search backend for AI agents: SearXNG meta-search + RAG with citations + GitHub project selection + Tavily-compatible API. Free & local.
MIT ★ 5 -
MCP server for self-hosted SearXNG: web/news search, readable-text page extraction, and search-and-extract in one call, with engine-fallback against silent zero-result responses
MIT ★ 1 -
Self-hosted flomo MCP server based on flomo web session credentials.
npm MIT -
spectrawlgithub.com/Pyx-Corp/spectrawlupdated this year
The unified web layer for AI agents. Search (8 engines), stealth browse, auth, and act on 24 platforms. One npm install, self-hosted.
MIT ★ 28 -
The web toolkit for AI coding agents — scrape, search, map, crawl, index, extract, diff, serve. Single static binary, MIT. Self-hosted Firecrawl/Exa/Tavily alternative.
MIT -
A simple, locally hosted Web Search MCP server for use with Local LLMs
npm MIT ★ 1,207 -
🔎 One query. Three search engines. Self-hosted web search for AI agents via HTTP API or MCP. Aggregates DuckDuckGo, Brave & Yahoo. Built with Rust. No API keys required.
NOASSERTION ★ 135 -
firescraplinggithub.com/grayaa/firescraplingupdated this year
Self-hosted cost-control layer for web scraping. Bring your own Scrape.do/Scrapfly key, escalate fetch tiers only when needed, get LLM-ready Markdown. MCP server included.
NOASSERTION -
Self-hosted search and research infrastructure for AI agents. Query web and specialist sources through MCP, run durable research jobs, and monitor results with replayable change events. SearXNG-compatible HTTP API.
MIT ★ 14 -
The open-source gateway to authenticated web data — connect to any user-authorized site via REST API, hosted link, or MCP, with MFA handling under a read-only posture.
MIT ★ 12 -
Self-hosted Web Bot starter for teams running OpenClaw and MCP knowledge servers.
Apache-2.0 -
ollama-web-tools-mcpgithub.com/chakkritte/ollama-web-tools-mcpupdated this year
Claude Code MCP server for web search and fetch. Supports Ollama's hosted API instance.
MIT ★ 2 -
Self-hosted MCP servers for ChatGPT plugins: read web pages and documents, and call HTTP APIs on Cloudflare Workers.
MIT ★ 6 -
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)
-
Self-hosted MCP relay and web dashboard
npm MIT -
MCP stdio server exposing Ollama's hosted web search and web fetch APIs as tools for any MCP client.
pypi MIT -
AgentReady REST API and Docker deployment for AI agent readiness scanning. Scan websites, APIs, ecommerce stores, marketplaces, MCP servers and agent services for agentic-web compatibility, discoverability, OpenAPI, llms.txt, structured data, security and AI crawler readiness. Easy to self-host.
Apache-2.0 -
Self-hosted MCP server on Cloudflare Workers: web search (keyword + semantic), Reddit, X, YouTube, Discord and page reading — one worker, one token, no LLM in the pipe
MIT -
Open-source, self-hosted, privacy-first web and product analytics on Cloudflare Workers, D1, and R2. Add one script tag to your site, then read results in a local console or ask your AI assistant (MCP). Run `vizoalica env add` to set it up.
npm MIT -
Self-hosted meal log: MCP server for Claude and ChatGPT plus a web dashboard
MIT ★ 1 -
Self-hosted Remote MCP server for isolated reasoning, review, and web research.
MIT -
@caikiji/mcp-searchgithub.com/caikiji/mcp-searchupdated this year
MCP server for web search powered by SearXNG — privacy-respecting self-hosted metasearch
npm MIT -
Firecrawl Web Scraping Serverno update data
Enable advanced web scraping, crawling, and content extraction capabilities for your agents. Perform deep research, batch scraping, and structured data extraction with automatic retries and rate limiting. Support both cloud and self-hosted deployments with seamless integration into popular MCP clients.
smithery -
playwright-web-fetchergithub.com/ZaZ712/playwright-web-fetcherupdated this year
Deploy a real browser to the cloud. AI-agent ready web fetching API — bypass Cloudflare, get clean content, persist sessions. curl-friendly, self-hosted, Docker-ready.
★ 2 -
evocrawl-mcpgithub.com/evocrawl/evocrawl-mcp-serverupdated this year
MCP server for Evocrawl — search, scrape, and interact with the web. Supports both cloud and self-hosted instances. Features include web search, scraping, page interaction, batch processing, and LLM-powered content analysis.
npm MIT -
docsearchgithub.com/ypngslo/docsearchupdated this year
Self-hosted document search: local ingestion + pgvector retrieval pipeline, web console, and a ChatGPT MCP connector
MIT -
kotojigithub.com/necorox-com/kotojiupdated this year
MCP-native, self-hosted hosting for AI-built web tools — git as the source of truth, per-project subdomains, in-browser editing, and per-user MCP/API access.
AGPL-3.0 -
easyseachgithub.com/yoligehude14753/easyseachupdated this year
Hosted web-search MCP with focused, multi-round, and deep-search Agent Skill workflows
MIT -
Declarative service registry, web UI and MCP server for orchestrating Docker services across heterogeneous hosts
Apache-2.0 -
Self-hosted, AI-native CRM for solo founders. CLI + Web + MCP over one SQLite DB.
MIT -
Self-hosted, platform-independent home inventory with a Web UI and first-class LLM/MCP support
MIT -
Resilient multi-provider web search & content extraction for Claude Code, DeepSeek Harness, and any MCP host — quota-aware failover, JS rendering, and a CLI.
Apache-2.0 ★ 2 -
Self-hosted social media publishing: web UI, HTTP API, MCP endpoint and CLI in a single service
★ 17 -
Web scraping API and hosted MCP server for AI agents. Read public pages as Markdown, HTML, links, or JSON and run bounded crawls with Cloudflare Browser Run.
AGPL-3.0 -
Self-hosted AI web mail for all your mailboxes: an LLM judges what matters across inboxes and spam, pushes only that, and the web view opens on what needs you, to read, file and answer there. MCP server included.
MIT -
distillgithub.com/afly007/distillupdated this year
Self-hosted RAG pipeline — ingest technical PDFs and web pages into a searchable knowledge base, exposed as MCP tools
MIT ★ 2 -
Agent-powered task, habit & reminder system — macOS desktop widget + iPhone web widget, self-hosted, MCP-integrated. MIT.
MIT -
Official MrScraper power for Kiro: hosted MCP server and skills for web scraping, web search, and account usage.
MIT -
Pry — self-hosted web scraping and browser automation with Cloudflare bypass, MCP server and x402 pay-per-call.
NOASSERTION -
Contrie Plugin for Claude and Codex: web data with evidence from one public page, through Contrie's hosted MCP server.
MIT