Neura Market
ConechoAI/openai-websearch-mcp logo

ConechoAI/openai-websearch-mcp

Free

提供 OpenAI `web_search` 内置工具的 Python MCP 服务器。

FreeFree tier
Type
Open Source
Visit Website GitHub (40 stars)

About ConechoAI/openai-websearch-mcp

OpenAI WebSearch MCP Server is an advanced MCP (Model Context Protocol) server that provides intelligent web search capabilities using OpenAI's reasoning models. It is designed for AI assistants that need up-to-date information with smart reasoning. The server supports multiple OpenAI models (gpt-5, gpt-5-mini, gpt-5-nano, o3, o4-mini), features smart effort control, multi-mode search (fast iterations or deep research), localized results via user location, and flexible configuration through environment variables. It can be easily installed via uvx for use with Claude Desktop, Cursor, or other MCP-compatible clients. The repository includes detailed parameter documentation for seamless integration.

Key Features

Supports latest OpenAI reasoning models (gpt-5, gpt-5-mini, gpt-5-nano, o3, o4-mini)
Smart reasoning_effort defaults based on use case (low/medium/high/minimal)
Multi-mode search: fast iterations with gpt-5-mini or deep research with gpt-5
Localized results with optional user_location parameter
Rich parameter documentation for easy integration
Flexible configuration via environment variables (OPENAI_API_KEY, OPENAI_DEFAULT_MODEL)

Pros & Cons

Pros
  • Open source and free to use (no additional tool cost)
  • Easy one-click installation for Claude Desktop using uvx
  • Supports multiple OpenAI reasoning models including cutting-edge gpt-5 series
  • Flexible configuration and environment variable support for deployment
  • Detailed usage examples and parameter documentation
Cons
  • Requires an OpenAI API key which may incur usage costs
  • Only works with OpenAI models (not compatible with other providers)
  • Functionality limited to web search (no other tool integration)

Best For

Quick search for latest news, developments, or information using natural languageDeep research with high reasoning effort and advanced models for comprehensive answersIntegration with AI assistants in Claude Desktop, Cursor, or Claude Code for real-time web queriesLocation-based search customization for localized results

FAQ

What models does openai-websearch-mcp support?
It supports OpenAI's latest reasoning models: gpt-5, gpt-5-mini, gpt-5-nano, o3, o4-mini, as well as gpt-4o and gpt-4o-mini. The default model is gpt-5-mini.
How do I install openai-websearch-mcp?
For Claude Desktop, use the one-click uvx command: uvx --with openai-websearch-mcp openai-websearch-mcp-install, after setting your OPENAI_API_KEY. For other clients, add the server configuration to your MCP settings as shown in the README.
Can I use this with Cursor or Claude Code?
Yes, the server can be configured for Cursor by adding the MCP server settings. Claude Code automatically detects servers configured for Claude Desktop.
What parameters does the openai_web_search tool accept?
It accepts input (required), model (default gpt-5-mini), reasoning_effort (low/medium/high/minimal, smart default), type (web_search_preview), search_context_size (low/medium/high), and user_location (optional object for localized results).