Content
<div align="center">
# Tool List
**Real-time Trending Topics for Claude / Cursor / Windsurf via MCP**
**Empower Your AI to Sense Real-time Chinese Internet Trends — Weibo Hot Search · Zhihu Hot List · Bilibili Hot · Baidu · Douyin · Toutiao · 36Kr**
[](LICENSE)
[](https://python.org)
[](https://modelcontextprotocol.io)
</div>
---
## What Problem Does This Tool Solve?
**Scenario 1 (Self-Media/Content Creators)**
> Manually open Weibo, Zhihu, Douyin, Baidu to check trending topics every day, switching between 5 apps to find today's writeable topics.
> → Install and ask Claude: "What are the trending topics today? Analyze them based on my account direction."
**Scenario 2 (AI Agent Developers)**
> Building a Chinese AI Agent requires real-time public opinion data, but existing MCP tools either rely on Node.js or have a single data source that may fail.
> → `pip install` and use, Python native, dual data sources, built-in cache.
**Scenario 3 (Claude / Cursor / Windsurf Users)**
> AI doesn't know what happened today, resulting in disconnected responses, requiring manual copying of trending topics.
> → Configure once, and then AI will fetch real-time data automatically.
---
## Comparison with Similar Tools
| | mcp-hotnews-server | **china-hot-mcp** |
|---|---|---|
| Language | Node.js / TypeScript | ✅ **Python Native** |
| Installation | `npx` (requires Node environment) | ✅ **`pip install`** |
| Tool Invocation | `get_hot_news([1,3,7])` numeric ID | ✅ **`weibo_trending()` semantic naming** |
| Data Source | Single vvhan.com | ✅ **Directly connect to platform API + backup source double insurance** |
| Cache Mechanism | None | ✅ **5-minute TTL, anti-flow limit** |
| Documentation | English | ✅ **Chinese priority** |
---
## Quick Start (5 minutes)
### Installation
```bash
pip install china-hot-mcp
```
### Configure Claude Desktop
`~/Library/Application Support/Claude/claude_desktop_config.json`
```json
{
"mcpServers": {
"china-hot": {
"command": "china-hot-mcp"
}
}
}
```
### Configure Cursor / Windsurf
```json
{
"mcpServers": {
"china-hot": {
"command": "python",
"args": ["-m", "china_hot_mcp.server"]
}
}
}
```
Restart the client to use.
---
## Available Tools (MCP Tools)
| Tool | Platform | Returned Content |
|------|------|---------|
| `weibo_trending(limit)` | Weibo | Real-time Hot Search List + Heat Value |
| `zhihu_trending(limit)` | Zhihu | Hot List Questions |
| `bilibili_trending(limit)` | Bilibili | Hot Videos + Play Count |
| `baidu_trending(limit)` | Baidu | Baidu Hot Search List |
| `douyin_trending(limit)` | Douyin | Hot Topics List |
| `toutiao_trending(limit)` | Toutiao | Hot Articles List |
| `kr36_trending(limit)` | 36Kr | Tech Entrepreneurship Hot Articles |
| `all_trending(limit)` | **All Platforms** | 5 Major Platforms Summary |
---
## Practical Usage Examples
**Content Creation Topic Selection:**
```
User: Help me analyze today's trending topics across all platforms and find 3 suitable topics for technology-related content.
Claude calls all_trending(limit=10), returns real-time data, and analyzes:
1. Zhihu Hot List #3 "How to view XXX technology breakthrough" → Suitable, professional interpretation angle
2. Weibo Hot Search #7 "XXX conference" → Suitable, product review angle
3. Toutiao Hot List #2 "AI + industry dynamics" → Suitable, trend analysis angle
```
**AI Agent Public Opinion Monitoring:**
```python
# In your Agent, Claude will automatically call tools when needed
result = agent.run("Monitor today's tech industry-related trending topics and alert if there are major events")
```
---
## Architecture Description
```
Request → Cache Check (5min TTL)
↓ Not hit
Directly connect to platform API (Weibo/Zhihu/Bilibili priority)
↓ Failed
Backup aggregated source (vvhan.com)
↓ Failed
Friendly error prompt
```
- **MCP Framework**: Anthropic official `mcp[cli]` Python SDK
- **HTTP Client**: `httpx` asynchronous request
- **Concurrency**: `asyncio.gather` multi-platform parallel fetching
---
## Local Development & Debugging
```bash
git clone https://github.com/EA-Studio-SHARK/china-hot-mcp
cd china-hot-mcp
pip install -e .
# MCP Inspector visual debugging
npx @modelcontextprotocol/inspector python -m china_hot_mcp.server
```
---
## Who Is It For?
- ️ **Self-Media/Content Creators** — AI topic selection assistant, goodbye to manually flipping through 5 apps daily
- 🤖 **AI Agent Developers** — Access real-time Chinese public opinion perception capabilities for Agents
- 📊 **Operations/Brand Monitoring** — Real-time tracking of industry hot words changes
- 🛠️ **Claude / Cursor / Windsurf Users** — Let AI know what happened today
---
## Related Projects
- [awesome-mcp-zh](https://github.com/EA-Studio-SHARK/awesome-mcp-zh) — MCP China Developer Practical Toolkit (精选 + 亲测 + 踩坑记录)
- [local-ai-china](https://github.com/EA-Studio-SHARK/local-ai-china) — Local AI one-click deployment guide (solve China network + API cost issues)
---
## Contact
- 📬 Telegram: [@ExploreAllStudio](https://t.me/ExploreAllStudio)
- 💼 Agent customization development / MCP integration cooperation welcome
---
<div align="center">
**Star ⭐ is the greatest support if you find it useful**
</div>
