mcpField
H
Hot Topics MCP

Hot Topics MCP

Remote
github.comBusiness & Marketing0Updated 4w ago

About Hot Topics MCP

Hot Topics MCP gathers trending topics and community discussions for content and marketing agents doing research. It combines sources from China, Google Trends, HackerNews, and StackExchange, with graceful handling when a source is unavailable.

Key features

  1. 01Collect trends from four topic sources
  2. 02Retrieve HackerNews and StackExchange threads
  3. 03Filter trends by content track
  4. 04Blacklist sensitive topics
  5. 05Handle unavailable sources without guesses

Use cases

  1. 01Research current topics before drafting content
  2. 02Find user pain points for GEO research
  3. 03Compare trends across China and US sources

MCP support

The MCP server exposes get_trending and get_reddit_intel over stdio. The tools return structured trend and community-thread data from the listed sources.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://hot.mcp.goaimoat.com/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

hot-topics-mcp

M8ven Score

Trending topics for AI agents — Weibo / Baidu / Zhihu (China) + Google Trends (US), plus HackerNews & StackExchange community intel. 2 tools. Zero dependencies, pure Python stdlib.

  • Hosted (no install): https://hot.mcp.goaimoat.com/mcp
  • Registry name: com.goaimoat/hot-topics
  • Free: 10 calls per day per IP. Then $0.03/call.

Built for content-creation and marketing agents that need to know what people are actually talking about right now before they write. Battle-tested in production driving a daily automated content pipeline.

Tools

ToolWhat it gives the agent
get_trendingLive hot topics from 4 sources, scored against content tracks (亲子教育, AI工具, EN_kids, EN_ai, all, raw), with a built-in sensitive-topic blacklist (death/disaster/crime/politics) so agents never draft around toxic trends
get_reddit_intelHigh-engagement community threads (HackerNews + StackExchange: Parenting/Workplace/Expatriates) for GEO (Generative Engine Optimization) research — the pain points real users are asking about this week

Why this exists

  • Trending data is scattered across platforms that expose no API, and each source covers a different audience. An agent that reads only one is working from a partial picture.
  • This server collects them into one structured response, so the agent sees the whole picture in a single call.
  • Multi-source with graceful degradation: any one source failing does not break the pipeline. A source that cannot be read is reported as unavailable rather than filled in with guesses.

Quick start

Requires Python 3.8+. No pip install needed.

Claude Desktop / WorkBuddy (mcp.json)
{
  "mcpServers": {
    "hot-topics": {
      "command": "python",
      "args": ["/absolute/path/to/server.py"]
    }
  }
}

Then ask your agent: "What are today's trending topics in the AI工具 track?" or "Get reddit intel for the kids track and draft a post."

Example tool calls
{"name": "get_trending", "arguments": {"track": "AI工具"}}
{"name": "get_trending", "arguments": {"track": "EN_ai", "sources": ["google_trends"]}}
{"name": "get_reddit_intel", "arguments": {"track": "poa"}}

Hosted API (no install, works behind any firewall)

Don't want to run it yourself, or your network can't reach these sources? A hosted version runs 24/7 on our Singapore server — one HTTP GET, JSON back:

GET http://43.160.199.215/hotapi/v1/hot?track=EN_ai&key=YOUR_KEY
GET http://43.160.199.215/hotapi/v1/reddit?track=poa&key=YOUR_KEY
GET http://43.160.199.215/hotapi/v1/health

30-day access key: $9 → https://niebingyu.gumroad.com/l/hot-topics-api

Notes

  • 10-minute in-memory cache per source; polite to upstreams.
  • All sources are public endpoints used for research; respect each platform's terms.
  • Protocol: MCP over stdio, JSON-RPC 2.0, protocol version 2024-11-05.

License

MIT

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.