brave_web_searchLABSCO SUMMARY
Brave's index, with a separate lane for places.
Two tools, cleanly divided: one for the general web, one for businesses and locations. The split matters because a query like 'coffee near Shoreditch' and a query about a library's release notes want different backends.
What it is
A Python MCP server that puts the Brave Search API in front of an MCP client, with a web search tool and a local-places search tool.
What you get
- Web search performed through the Brave Search API and returned to the session
- Local search for businesses and places, separate from the general web index
Requirements
A Brave Search API key in BRAVE_API_KEY, and Python 3.11+ with uv.
Setup effort
Build from source
USE CASES
01
Getting current facts into a conversation
Run a web search through Brave when the answer needs to be from this week rather than from training data.
02
Finding a business or a place
Use the local search lane, which returns businesses and places rather than pages about them.
03
Checking a claim before you repeat it
Search for the primary source rather than accepting a summary, and read what comes back.
2 tools
Listed in project docs. Your client may show a different list.
brave_local_searchBEFORE YOU INSTALL
What to know first
Tools you get
2 tools named in the project's documentation.
Requires
- uv on PATH
- Python 3.11+
- BRAVE_API_KEY in your environment
INSTALLATION
Connect your client
MCP Clients
Claude Desktop
CLAUDE DESKTOP · TERMINAL
git clone https://github.com/thomasvan/mcp-brave-search.git
cd mcp-brave-search
uv venv
uv pip install -r requirements.txtThen add this to claude_desktop_config.json:
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"brave-search": {
"command": "uv",
"args": [
"--directory",
"<YOUR_PATH_TO_MCP_PYTHON_MCP_BRAVE>",
"run",
"server.py"
],
"env": {
"BRAVE_API_KEY": "<YOUR_BRAVE_API_KEY>"
}
}
}
}The path ending in mcp-brave-search\src is the full path of the src folder inside the mcp-brave-search folder the clone created (there is no mcp-python folder).
