Allows web search using Google Search engine via Serper API, supporting parallel multi-query searching and advanced search syntax.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@MiniMax Search MCP Server Search for 'latest AI news' and 'machine learning trends' simultaneously"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MiniMax Search is an MCP (Model Context Protocol) server that provides web search and browsing capabilities.
This project uses the standard MCP Server protocol, compliant with MCP specifications:
ā
Complete list_tools() implementation
ā
Complete call_tool() implementation
ā
Standard stdio_server() communication
Related MCP server: crawl-mcp
š Multi-Engine Search: Supports Google search engine
š Parallel Search: Native support for parallel multi-query search
š Batch Browsing: Support for batch browsing of multiple URLs
š¤ Intelligent Understanding: Uses MiniMax LLM to understand web content and answer questions
šÆ Advanced Search: Supports Google advanced search syntax
š Auto Fallback: Automatically switches to other engines when search fails
# Run directly from Git repository
uvx --from git+ssh://git@github.com:MiniMax-AI/minimax_search.git minimax-search # Run from local directory
uvx --from /xxx/minimax_search minimax-searchAdd to your MCP client configuration file (e.g., mcp.json):
{
"mcpServers": {
"minimax_search": {
"command": "uvx",
"args": [
"--from",
"git+ssh://git@github.com:MiniMax-AI/minimax_search.git",
"minimax-search"
],
"env": {
"MINIMAX_API_KEY": "your_minimax_api_key",
"SERPER_API_KEY": "your_serper_api_key",
"JINA_API_KEY": "your_jina_api_key"
}
}
}
}Search multiple queries simultaneously, returning brief results (title, URL, snippet).
Parameters:
queries (array of strings, required): List of queries, supports Google advanced search syntax
Supported Search Engines:
Google Search (via Serper API)
Advanced Search Syntax:
site:example.com - Limit to specific site
intitle:keyword - Title contains keyword
inurl:keyword - URL contains keyword
"exact match" - Exact phrase match
Example:
{
"queries": ["Python asyncio tutorial", "Python threading vs asyncio"]
}Visit multiple web pages, use MiniMax LLM to understand content and answer questions.
Parameters:
urls (array of strings, required): List of target web page URLs
query (string, required): Question to be answered
Example:
{
"urls": [
"https://docs.python.org/3/library/asyncio.html",
"https://realpython.com/async-io-python/"
],
"query": "Summarize the main features and use cases of asyncio"
}Basic Search Functionality:
SERPER_API_KEY: Google Search
Get it from: https://serper.dev/
Free tier: 2,500 requests/month
Web Browsing Functionality:
JINA_API_KEY: Web content reading
Get it from: https://jina.ai/
Free tier available
Browse Functionality (LLM Understanding):
MINIMAX_API_KEY: MiniMax API Key
Get it from: https://platform.minimax.io/
Once the MCP server is started, the Agent can use the following tools:
Parallel search for multiple queries:
User: Search for "Python asyncio" and "Python threading" differences simultaneously
Agent: [Calls search tool]
ā search(queries=["Python asyncio", "Python threading"])
ā Returns search results for both queries (executed in parallel)Batch browse multiple web pages:
User: Visit these links and summarize the main features of asyncio
- https://docs.python.org/3/library/asyncio.html
- https://realpython.com/async-io-python/
Agent: [Calls browse tool]
ā browse(
urls=["https://docs.python.org/...", "https://realpython.com/..."],
query="Summarize the main features of asyncio"
)
ā Returns comprehensive summary and answerminimax_search/
āāā server.py # MCP Server entry point (2 tools)
āāā minimax_search_browse.py # Core search and browse implementation
āāā pyproject.toml # Project configuration
āāā README.mdParallel Search:
Native support for queries array
Concurrent execution using thread pool
Automatic formatting and grouping of results
Batch Browsing:
Native support for urls array
Jina Reader extracts web content (converts to Markdown)
MiniMax LLM understands content and generates comprehensive answers
MIT
This project is based on the MiniMax-M2 Model project
Explore specific information in a list of urls. The parameters are a url list and a query. The urls will be browsed, and each content will be sent to a Large Language Model (LLM) as the based information to answer the query.
| Name | Required | Description | Default |
|---|---|---|---|
| urls | Yes | The url list. | |
| query | Yes | The query. A detailed natural language query is recommended. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description discloses that URLs are browsed and content is sent to an LLM to answer the query, which is a key behavioral trait. However, it lacks details on side effects, permissions, rate limits, or privacy implications, which are needed given no annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise with two sentences that front-load the purpose and then explain the operation. No redundant or extraneous information is present.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The description does not explain the output format or return value, nor does it address potential errors, timeouts, or limitations. Given the lack of an output schema, this is a significant omission for a tool that fetches URLs.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, and the description adds a recommendation for a detailed natural language query. However, the description largely restates the parameter names without adding significant new meaning beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool explores specific information from a list of URLs using a query and an LLM. However, it does not explicitly differentiate the tool from its sibling 'search', which may lead to ambiguity.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance is provided on when to use this tool versus alternatives, nor any conditions or prerequisites. The description simply explains the tool's operation without contextual usage advice.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Web search in parallel. The parameter is a list of queries. The queries will be sent to a search engine. You will get the brief search results with (title, url, snippet)s for each query.
| Name | Required | Description | Default |
|---|---|---|---|
| queries | Yes | The queries. Google advanced search operators are supported. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full burden. It discloses that queries are sent to a search engine and results are returned, but does not state whether the tool is read-only, rate limits, error handling, or what happens with failed queries. The read-only nature is implied but not explicit.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise with three sentences that are front-loaded. Every sentence adds useful context without unnecessary detail. No redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a single-parameter search tool with no output schema, the description adequately explains input (list of queries) and output (brief results with title, url, snippet). It lacks details on error handling or limits, but the tool's complexity is low, so completeness is sufficient.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% with a description that already mentions Google advanced operators. The tool description adds value by explaining parallel execution and output format (title, url, snippet), but does not significantly extend parameter meaning beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states it performs web search in parallel using a list of queries and returns brief results with title, url, snippet. It implicitly distinguishes from sibling tool 'browse' through the verb 'search', but no explicit differentiation is provided.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description mentions parallel execution for multiple queries, implying batch use, but lacks explicit guidance on when to use this tool vs. the sibling 'browse' tool or alternative contexts. No exclusions or prerequisites are mentioned.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v1.0.0browsesearchThe two tools have clearly distinct purposes: search finds web results, while browse extracts information from specific URLs. No overlap in functionality.
Both tool names are single verbs ('browse', 'search') following a uniform imperative style. No mixed conventions or irregularities.
With only 2 tools covering web search and targeted URL exploration, the count is minimal yet perfectly scoped for the server's purpose.
The server provides a complete surface for web search and browsing: search retrieves initial results, and browse extracts detailed information from selected URLs. No obvious gaps.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
MCP server for MiniMax H3 multimodal video generation
Live AI-native web search with citations. One tool for every MCP client. Flat per-request pricing.
MCP server for building and testing AI agents with multi-model experimentation and insights.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/MiniMax-AI/minimax_search'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
