About Google Search MCP Server
Google Search MCP Server (Nimrod) gives Claude web research tools, including quality-scored Google search, clean page extraction and multi-source topic research. It suits users whose agents need usable web evidence.
Key features
- 01Google search with quality scores
- 02Webpage conversion to clean markdown
- 03Concurrent extraction of up to five URLs
- 04Single-call search, ranking and synthesis workflow
- 05Hosted remote service and local desktop version
Use cases
- 01Retrieving and filtering web sources for Claude
- 02Extracting the main content of specific web pages
- 03Researching a topic across multiple sources
MCP support
This is an MCP server that exposes google_search, extract_webpage_content, extract_multiple_webpages and research_topic tools to MCP clients such as Claude. It can be connected through a remote HTTP endpoint or a local installation.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project README
Nimrod
Web research your agent can actually use.
Quality-scored Google search, clean webpage extraction, and deep multi-source
research for Claude — over the Model Context Protocol.
Website · Docs · Pricing · Portal
What it is
Nimrod gives Claude four research tools:
| Tool | What it does | Cost |
|---|---|---|
google_search | Organic Google results with quality scores, authority ratings, and source-type classification | 1 credit |
extract_webpage_content | Any URL → clean markdown, boilerplate stripped | free |
extract_multiple_webpages | Up to 5 URLs, fetched concurrently | free |
research_topic | Search → rank → dedupe → extract → synthesis, in one call | 1 credit + 1 per focus area |
Why not just let the model search? Because raw search is where agent runs go to die: Nimrod returns ranked evidence instead of SEO sludge, as token-efficient markdown instead of raw HTML, with deterministic costs (1 credit = 1 search) and date/site/language control.
New accounts start with a free 50-credit trial — no card required.
Get connected
Claude.ai & Claude Desktop (zero install): Settings → Connectors → Add
custom connector → paste https://nimrod.orchis.ai/mcp → sign in with your
email. Done.
Claude Code (hosted):
claude mcp add nimrod --transport http https://nimrod.orchis.ai/mcp \
--header "Authorization: Bearer nmk_your_key"
Nimrod Desktop (full toolkit — recommended for power users): a locally installed MCP server with six research skills, a deep-research agent, and search-coach hooks (Claude Code features). Extraction is free on every plan; on Desktop it also runs on your own machine, so the pages you read stay private. Grab it from Releases:
- Windows —
nimrod.mcpbfor Claude Desktop (double-click, paste key), ornimrod-desktop.exefor Claude Code:nimrod-desktop setup, thenclaude mcp add nimrod -- path\to\nimrod-desktop.exe - macOS —
nimrod-desktop-macos(universal: Apple Silicon + Intel):chmod +xit, then a one-timexattr -d com.apple.quarantine nimrod-desktop-macos(not yet Apple-signed), thensetupandclaude mcp addas above - Linux —
nimrod-desktop-linux:chmod +x,setup,claude mcp add
In the MCP Registry
Nimrod is listed in the official MCP Registry
as ai.orchis/nimrod — MCP clients that browse the registry (including
VS Code's MCP gallery) can discover and install it from there.
Built to pair
Run Nimrod beside your other MCP tools — like Konnect for KiCAD: live parts data from your EDA workflow, live web research from Nimrod, one agent, no stale training cutoff.
Support
Docs: nimrod.orchis.ai/docs.html · Email: [email protected] · Terms · Privacy
The original open-source BYOK server this project grew from is preserved
on the historical-oss branch (unmaintained).
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.