Skip to content

searxng-mcp

SearXNG search MCP for Claude Code and Grok — privacy-respecting web search and page extraction for AI agents and any MCP client.

plugin-validate License: FSL-1.1-ALv2 Claude Code plugin GitHub Docs

Connect your SearXNG instance to an LLM. You get search, multi-query research, and readable page extraction as MCP tools. Model-visible output stays short; full payloads live in hidden _meta.

Install

Claude Code plugin (recommended):

/plugin marketplace add 88plug/claude-code-plugins
/plugin install searxng@88plug

Grok Build

grok plugin marketplace add 88plug/claude-code-plugins
grok plugin install searxng@88plug --trust

Standalone, any MCP client (no install step):

uvx --from git+https://github.com/88plug/searxng-mcp searxng-mcp

Note

You need a reachable SearXNG instance. Default backend is http://127.0.0.1:8890. Point elsewhere with SEARXNG_MCP_BASE_URL.

Quickstart

export SEARXNG_MCP_BASE_URL=http://127.0.0.1:8890
uvx --from git+https://github.com/88plug/searxng-mcp searxng-mcp

In your MCP client, call health. A healthy backend returns ok: true with backend, cache, and render details. Then call search with a query for a compact ranked list (full payload in _meta).

Rendered fetch ships in the default install. Local Chromium/Chrome is used when present; otherwise the first rendered fetch bootstraps Playwright Chromium into the user cache.

MCP tools

Tool Use when
search One query, compact ranked hits
search_many Parallel query variants, merged ranking
search_and_fetch One query plus source extraction
research Multi-query investigation with citations
fetch_url Read one known URL
fetch_many Read several URLs in parallel
health Backend / cache / render readiness

Tools surface as mcp__searxng__search, mcp__searxng__fetch_url, and so on.

Also exposed:

  • Resources: searxng://config, searxng://guide
  • Prompts (compatibility): quick_lookup, deep_research, research_workflow

Full parameter tables live in the Tool Reference.

Why this shape

Capability Detail
Token-efficient output Compact model-facing text; full raw data in _meta
Research workflows Fan-out search, merge/dedupe, fetch with citations
Rendered pages Playwright auto-fallback for JS-heavy HTML
Dual transport stdio for local clients; streamable-http for shared deploy
Self-hosted Docker + hardened Compose; your SearXNG, your network

Where next

Features

Feature Detail
Token-efficient MCP tools Compact model-visible output; full payloads in hidden _meta
Parallel research Multi-query search and fetch fan-out with dedupe and merged ranking
Rendered extraction Playwright/Chromium path for JS-heavy pages; no extra install flags
Self-hosted deployment stdio, streamable-http, SSE, Docker, and Compose for private MCP services
Thin by design SearXNG does the search; this server shapes tools, cache, extract, transport

Development

From a checkout:

uv sync
uv run searxng-mcp

Build and test:

uv sync --all-groups
uv run pytest -q
uv run python -m compileall src
uv run mkdocs build --strict

Contributions welcome. See CONTRIBUTING.md and the code of conduct.