codex-radar-mcp
This MCP server exposes a set of tools that fetch real-time model IQ and recommendation data from codexradar.com, allowing MCP clients like MaiBot or Claude Desktop to answer questions about which AI model is currently the smartest, how to configure effort tiers for different usage scenarios, and whether any model has recently experienced a performance drop.
README
📡 Codex Radar MCP
把 codexradar.com 的实时模型智商数据,封装成标准 MCP(Model Context Protocol)服务器,让 MaiBot、Claude Desktop、Cursor 等 AI 助手能随时回答——"现在哪个模型最聪明?该换哪个档?有没有降智?"
✨ 特性
- 零配置数据源:直接抓取 codexradar.com 的公开 JSON 接口,无需登录、无需 API Key、不解析 HTML
- 5 个现成工具:综合智能 IQ、软件工程 IQ、视觉空间 IQ、社区体感分、订阅人数
- 分场景档位推荐:自动给出「日常开发 / 难题攻坚 / 后台自动化 / 跑龙虾」各自的模型 + 推理档位建议(即雷达站上的"建议换 max/xhigh/low")
- 开箱即用:
pip install+ 一行启动,Streamable HTTP 传输,主流 MCP 客户端通吃 - 自带缓存:120s 本地缓存,避免高频查询打爆源站
🏗️ 架构
codexradar.com(公开 JSON 接口,约 5 分钟刷新)
│ HTTPS
▼
codex-radar-mcp(FastMCP · Streamable HTTP · 0.0.0.0:8765/mcp)
│ 120s 本地缓存 / 中文档位对照 / 数据清洗
▼
MCP 客户端(MaiBot · Claude Desktop · Cursor · 任意 MCP Client)
🚀 快速开始
1. 安装
git clone https://github.com/MorphieEndless/codex-radar-mcp.git
cd codex-radar-mcp
python3 -m venv venv
venv/bin/pip install -r requirements.txt
2. 启动
./run.sh
# 或直接:venv/bin/python -u src/codex_radar_mcp.py
默认监听 0.0.0.0:8765,MCP 端点:http://127.0.0.1:8765/mcp
3. 健康检查
./healthcheck.sh http://127.0.0.1:8765/mcp
# 期望输出:✅ healthy: codex-radar 1.29.0
4. 接入 MCP 客户端
以 Claude Desktop 为例(claude_desktop_config.json):
{
"mcpServers": {
"codex-radar": {
"url": "http://127.0.0.1:8765/mcp",
"transport": "streamable_http"
}
}
}
5. 试试看
连上后直接问:
- "现在哪个模型综合 IQ 最高?"
- "日常开发用哪个模型的哪个档位最划算?"
- "gpt-5.6 最近有没有降智?"
🔧 环境变量
| 变量 | 默认值 | 说明 |
|---|---|---|
CODEX_RADAR_HOST |
0.0.0.0 |
监听地址(容器/远程访问需保持非 127.0.0.1) |
CODEX_RADAR_PORT |
8765 |
监听端口 |
CODEX_RADAR_TIMEOUT |
15 |
抓取远端超时(秒) |
CODEX_RADAR_CACHE_TTL |
120 |
本地缓存时间(秒) |
CODEX_RADAR_PROGRESS_ENABLED |
false |
是否发送 MCP progress notification;默认关闭,且不保证 MaiBot 将其渲染为群消息 |
🛠️ 工具清单
| 工具 | 参数 | 说明 |
|---|---|---|
get_model_ratings |
无 | 近 24h 社区体感评分(按模型分组) |
get_radar_insights |
无 | ★核心:综合 IQ + 4 场景档位推荐 |
get_intelligence_effort |
model? |
软件工程 IQ 明细(可过滤,如 gpt-5.6-sol) |
get_visual_spatial |
model? |
视觉空间 IQ(庞贝邻接 F1,可过滤) |
get_subscriber_count |
无 | 雷达订阅人数 |
📊 示例输出(真实数据)
调用 get_radar_insights 返回(节选):
{
"ok": true,
"generated_at": "2026-08-22T11:56:08+00:00",
"recommendations": [
{
"scene": "daily_development",
"scene_cn": "日常开发",
"items": [
{"model": "gpt-5.6-sol", "effort": "high", "iq": 94.54, "average_cost_usd": 4.167},
{"model": "gpt-5.5", "effort": "xhigh", "iq": 93.08, "average_cost_usd": 5.290}
]
},
{
"scene": "hard_problems",
"scene_cn": "难题攻坚",
"items": [
{"model": "gpt-5.6-sol", "effort": "ultra", "iq": 105.1, "average_cost_usd": 22.52}
]
}
]
}
🤖 接入 MaiBot(麦麦)
本项目最初就是为 MaiBot 定制的。MaiBot 通过 MCPBridge 插件接入:
- 编辑 MCPBridge 插件的
config.toml,在[servers]段的claude_config_json中填入:
{
"mcpServers": {
"codex-radar": {
"url": "http://172.17.0.1:8765/mcp",
"transport": "streamable_http",
"enabled": true
}
}
}
172.17.0.1是 Docker 容器访问宿主机的 docker0 网关地址,按你的实际网络调整。
- 在群里对麦麦发
/mcp reconnect(或重启 MaiBot 容器) - 发
/mcp tools确认 5 个工具已列出
6. 测试
安装开发依赖并运行单元测试:
python3 -m venv venv
venv/bin/pip install -r requirements-dev.txt
venv/bin/pytest -q
测试覆盖模型过滤、缓存命中、并发 single-flight、上游失败重试和字段缺失时的输出。
⚠️ 免责声明
- 本项目的全部数据来自 codexradar.com 的公开接口,数据版权归源站所有,本项目仅做转发与格式整理。
- 本项目与 codexradar.com 官方无任何隶属/合作关系。
- 请合理设置
CODEX_RADAR_CACHE_TTL,不要高频请求源站。 - 数据仅供参考,不构成任何决策依据。
📄 License
MIT © 2026 MorphieEndless
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.