searchLABSCO SUMMARY
Three search engines, chosen for where you are sitting.
The README's own comparison is the useful part: Brave is high quality but has a demanding signup, Metaso is free and summarises but is slow and unofficial, Bocha is fast and good but paid. Decide once per deployment — the server runs a single engine at a time.
What it is
A Python search server with three interchangeable back ends: Brave Search, Metaso and Bocha. One is active at a time, chosen by configuration.
What you get
- search runs a web query against whichever engine is active
- With Brave: paging and offset control, plus location_search for places and businesses, which falls back to web search when it finds nothing
- With Metaso: a concise or in-depth answer mode, and scholar_search for academic material
- With Bocha: a freshness filter for the past day, week, month or year, an optional detailed summary, result counts, paging and related images
- Results carry the title, URL, source, snippet and publication time, alongside result and page totals
- Each engine is a self-contained module you can lift out and use elsewhere
Requirements
Python, and a credential for whichever engine you enable: an API key for Brave or Bocha, or session cookies for Metaso. The Metaso path additionally needs a Chromium browser installed for its driver. Only one engine is active per run.
Setup effort
Build from source
USE CASES
01
Getting current information into an answer
Let the client search on your behalf and work from the snippets rather than from memory.
02
Finding a place or a business
Run a location query on the Brave back end, which falls back to ordinary web search when nothing matches.
03
Looking for papers rather than blog posts
Use the academic search on the Metaso back end, in concise or in-depth mode.
04
Restricting results to recent coverage
On the Bocha back end, filter to the past day, week, month or year instead of putting a date in the query.
3 tools
Listed in project docs. Your client may show a different list.
location_searchscholar_searchBEFORE YOU INSTALL
What to know first
Tools you get
3 tools named in the project's documentation.
Requires
- Python 3.10+
- uv 0.24.0 or later
- One engine's credential: BRAVE_API_KEY (Brave), BOCHA_API_KEY (Bocha), or METASO_UID and METASO_SID (Metaso)
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
You clone this one yourself.
Server setup
TERMINAL
git clone https://github.com/fengin/search-server.gitMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"search": {
"command": "uv",
"args": [
"--directory",
"<YOUR_PATH_TO_SEARCH_SERVER>",
"run",
"search"
],
"env": {
"SEARCH_ENGINE": "brave",
"BRAVE_API_KEY": "<YOUR_BRAVE_API_KEY>"
}
}
}
}This block turns on the Brave engine only; <YOUR_BRAVE_API_KEY> is a Brave Search API key from api-dashboard.search.brave.com. SEARCH_ENGINE also takes bocha, with BOCHA_API_KEY (open.bochaai.com), or metaso, with METASO_UID and METASO_SID; Metaso also needs pip install playwright>=1.35.0 and playwright install chromium.
