# leehanchung/bing-search-mcp [Health: Active]

**Category:** 🔎 Search & Data Extraction  
**Repository:** https://github.com/leehanchung/bing-search-mcp  
**GitHub Stars:** 79  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/leehanchung-bing-search-mcp

## Description
Web search capabilities using Microsoft Bing Search API

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `uvx` (confidence: high):

```json
"mcpServers": {
  "bing-search-mcp": {
    "command": "uvx",
    "args": ["bing-search-mcp"],
    "env": {
      "BING_API_KEY": ""
    }
  }
}
```

**Requires environment variables:** `BING_API_KEY` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What leehanchung/bing-search-mcp MCP server does

The leehanchung/bing-search-mcp MCP server connects an MCP-compatible client to Microsoft Bing Search API. It gives an AI assistant three search operations: general web search, news search, and image search. This makes it suitable for agents that need to find websites and general information, look up recent events, or locate visual content through Bing.

The server is intended for local execution and requires Python 3.10 or newer, a Microsoft Bing Search API key, and an MCP client. The README specifically identifies Claude Desktop and Cursor as compatible client examples.

## How it works

The client sends a tool call to the local MCP process. The server forwards the request to Bing Search API and returns the search response to the client. Web searches accept a query, result count, offset, and market. News searches accept a query, count, market, and freshness value. Image searches accept a query, count, and market.

The available operations are:

- `bing_web_search` for general web results, with a default count of 10 and offset of 0.
- `bing_news_search` for news articles and current events, with a default count of 10 and a default freshness of `Day`.
- `bing_image_search` for image results, with a default count of 10.

Rate limiting is included to help prevent API abuse, and the project documents error handling as part of the server behavior.

## Setup and configuration

To install from the source repository, create a virtual environment with `uv venv`, activate it, and install the project in editable mode with `uv pip install -e .`. The documented runnable command is `uvx bing-search-mcp`.

Set `BING_API_KEY` before starting the process. `BING_API_URL` can also be set when a non-default endpoint is needed; the documented default is `https://api.bing.microsoft.com/`. The API key is obtained by creating a Bing Search resource in the Microsoft Azure Portal and retrieving its key from the resource's Keys and Endpoint section.

For Claude Desktop, add a server entry whose command is `uvx`, pass the local `bing-search-mcp` package or executable in the arguments, and provide `BING_API_KEY` in the entry's environment. Keep the key outside prompts and source-controlled configuration where possible.

## Tools and capabilities

The leehanchung/bing-search-mcp MCP server separates Bing's supported search modes instead of combining them into one tool. Use web search for broad information retrieval, news search when recency matters, and image search for visual-content discovery. Market parameters such as `en-US` allow the request to specify a regional search setting. News requests also expose a freshness parameter, with `Day` shown as the default.

This design fits agents that need to choose a search mode explicitly and pass basic pagination or localization options. It does not describe tools for opening pages, extracting page contents, downloading images, or performing browser actions.

## Limitations and notes

A Bing Search API key is required, so the server cannot run without credentials. API availability, quotas, and charges depend on the Microsoft Bing Search resource used; the provided material does not specify a pricing tier or quota. The README documents search requests and their parameters but does not state that the server crawls pages, summarizes results, or returns full page content.

The project is released under the MIT License. Its documented client setup covers local MCP use; no hosted endpoint is provided in the supplied material.

_Full upstream README: https://allmcps.com/mcp/leehanchung-bing-search-mcp/readme_