mcpField
S
Searchinfinity Mcp

Searchinfinity Mcp

Local
github.com/volcengineSearch333Updated 2w ago

About Searchinfinity Mcp

Searchinfinity Mcp connects MCP clients to Volcengine's web search API, including Doubao Search Custom and basic image search. It provides a web_search tool for retrieving current external search results and supports API key or AK/SK authentication.

Key features

  1. 01Web search through web_search
  2. 02Basic image search
  3. 03API key or AK/SK authentication
  4. 04stdio, SSE, and streamable HTTP modes

Use cases

  1. 01Find current information for an AI workflow
  2. 02Search for travel information such as Beijing-area trips

MCP support

Exposes the web_search tool to MCP clients for web and basic image searches through Volcengine's API. Clients can connect using stdio, SSE, or streamable HTTP.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"mcp_server_askecho_search_infinity": {
"command": "uvx",
"args": [
"git+https://github.com/volcengine/mcp-server#subdirectory"
],
"env": {
"VOLCENGINE_ACCESS_KEY": "<VOLCENGINE_ACCESS_KEY>",
"VOLCENGINE_SECRET_KEY": "<VOLCENGINE_SECRET_KEY>",
"ASK_ECHO_SEARCH_INFINITY_API_KEY": "<ASK_ECHO_SEARCH_INFINITY_API_KEY>"
}
}
}
}
Run commanduvx 'git+https://github.com/volcengine/mcp-server#subdirectory'
Environment variables to set
VOLCENGINE_ACCESS_KEYSecret
VOLCENGINE_SECRET_KEYSecret
ASK_ECHO_SEARCH_INFINITY_API_KEYSecret

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
pypigit+https://github.com/volcengine/mcp-server#subdirectory

Tools

1

Saved tool list · updated 2w ago

ParameterTypeRequiredDescription
AuthLevelintegerNo
CountintegerNo
QuerystringYes
SearchTypestringNo
TimeRangestringNo

Details

From the project README

联网搜索API MCP Server

版本信息

v0.2.0

产品描述

火山引擎联网搜索API,提供豆包搜索 Custom 版网页与基础图片搜索能力,帮助大模型获取更准确、更新鲜的外部信息。

分类

火山引擎云原生

标签

  • 搜索工具
  • 知识获取

Tools

本 MCP Server 产品提供以下 Tools (工具/能力):

Tool 1: web_search
类型

saas

详细描述

根据用户输入问题,返回联网搜索结果,支持网页与基础图片搜索。

调试所需的输入参数:

输入:

{
    "inputSchema": {
        "type": "object",
        "required": [
            "Query"
        ],
        "properties": {
            "Query": {
                "description": "用户搜索 query,1~100 个字符",
                "type": "string"
            },
            "Count": {
                "description": "返回条数;web 默认 10 条、最多 50 条,image 默认且最多 5 条",
                "type": "number"
            },
            "SearchType": {
                "description": "搜索类型,仅支持 web 或 image,默认 web",
                "type": "string"
            },
            "TimeRange": {
                "description": "web 搜索时间范围,可选 OneDay/OneWeek/OneMonth/OneYear 或 YYYY-MM-DD..YYYY-MM-DD",
                "type": "string"
            },
            "AuthLevel": {
                "description": "权威等级过滤,0 为默认,1 为非常权威",
                "type": "number"
            },
            "NeedContent": {
                "description": "是否返回网页正文,仅支持 web",
                "type": "boolean"
            },
            "NeedUrl": {
                "description": "是否返回网页 URL,仅支持 web",
                "type": "boolean"
            },
            "Sites": {
                "description": "仅搜索指定站点,以 | 分隔,最多 20 个,仅支持 web",
                "type": "string"
            },
            "BlockHosts": {
                "description": "排除指定站点,以 | 分隔,最多 5 个,仅支持 web",
                "type": "string"
            },
            "Industry": {
                "description": "行业搜索,可选 finance、game 或 gov,仅支持 web",
                "type": "string"
            },
            "QueryRewrite": {
                "description": "是否启用查询改写",
                "type": "boolean"
            },
            "ContentFormats": {
                "description": "网页正文格式,可选 text 或 markdown,仅支持 web",
                "type": "string"
            }
        }
    },
    "name": "web_search",
    "description": "联网搜索 API 调用"
}

输出:

联网搜索结果,结构参考官方 API 文档 https://www.volcengine.com/docs/87772/2272953
最容易被唤起的 Prompt示例

联网搜索北京周边游攻略

可适配平台

Claude Code, Codex, Trae,Cursor等支持MCP协议的AI工具

服务开通链接 (整体产品)

登录火山控制台,开通【联网搜索API】,服务开通链接:https://console.volcengine.com/search-infinity/web-search

API Key 创建链接:https://console.volcengine.com/search-infinity/api-key

鉴权方式

  • API Key鉴权
  • 火山引擎的AKSK鉴权体系

安装部署

前置准备
  • Python 3.12 / 3.13
  • UV Linux/macOS:
curl -LsSf https://astral.sh/uv/install.sh | sh

Windows:

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
安装

克隆仓库:

git clone [email protected]:volcengine/mcp-server.git
使用方法

启动服务器: UV

cd mcp-server/server/mcp_server_askecho_search_infinity
uv run mcp-server-askecho-search-infinity
# 使用sse/streamable-http模式启动(默认为stdio)
uv run mcp-server-askecho-search-infinity -t sse
uv run mcp-server-askecho-search-infinity -t streamable-http

部署

UVX

鉴权信息,火山引擎 AK/SK 与 ASK_ECHO_SEARCH_INFINITY_API_KEY 二选一即可

推荐:从 PyPI 或已配置的火山镜像源安装

该方式不依赖 GitHub,>=0.2.0 会自动使用可获取的最新发布版本。

{
  "mcpServers": {
    "mcp-server-askecho-search-infinity": {
      "command": "uvx",
      "args": [
        "--from",
        "mcp-server-askecho-search-infinity>=0.2.0",
        "mcp-server-askecho-search-infinity"
      ],
      "env": {
        "VOLCENGINE_ACCESS_KEY": "",
        "VOLCENGINE_SECRET_KEY": "",
        "ASK_ECHO_SEARCH_INFINITY_API_KEY": ""
      }
    }
  }
}
备用:从 GitHub 源码安装
{
  "mcpServers": {
    "mcp-server-askecho-search-infinity": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/volcengine/mcp-server#subdirectory=server/mcp_server_askecho_search_infinity",
        "mcp-server-askecho-search-infinity"
      ],
      "env": {
        "VOLCENGINE_ACCESS_KEY": "",
        "VOLCENGINE_SECRET_KEY": "",
        "ASK_ECHO_SEARCH_INFINITY_API_KEY": ""
      }
    }
  }
}

License

volcengine/mcp-server is licensed under the MIT License

Comments

Markdown supported

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.