AIHOT

一个自己找热点、自己写日报的网站框架。把信源和精选标准换成你的,它就是你的行业热点站。

What is AIHOT?

一个自己找热点、自己写日报的网站框架。把信源和精选标准换成你的,它就是你的行业热点站。

Overview of AIHOT

What it is

AIHOT is an open-source website framework that automatically collects materials from a set of sources, uses a large language model to filter and score them, selects noteworthy items, writes Chinese titles and summaries, clusters reports about the same event, calculates trending topics, and generates a daily newsletter. It is designed to be customized for any industry by replacing the sources and selection criteria.

Key features

  • Supports six types of sources: RSS, web page lists, JSON APIs, X accounts, WeChat public accounts, and custom script-pushed content.
  • Selection process: pre-filtering, two independent scoring rounds against the same criteria, and entry thresholds based on source tier.
  • Writing: generates Chinese titles, answer-first summaries, recommendation reasons, full-text translation for foreign articles, and extracts categories, tags, and facts.
  • Clustering: groups reports about the same event from different sources, with follow-up developments attached under the same event.
  • Trending: calculates heat per event based on the number of independent sources, with decay over 48 hours and indicators for rising or new topics.
  • Automated newsletters: daily, weekly, and monthly editions compiled from selected items and previous newsletters.
  • Admin panel: source management, content diagnostics, manual scoring corrections, selection evaluation, per-step model configuration, budget circuit breaker, and operation logs.

Use cases

  • Creating an industry-specific trending news website (e.g., for legal, HR, finance, precious metals).
  • Building a personalized news aggregator that filters and summarizes content according to custom criteria.
  • Generating automated daily briefings for internal teams or communities.

Getting started

bashgit clone https://github.com/KKKKhazix/AIHOT.git myhotcd myhotnode scripts/init-env.ts --llm-keydocker compose up -d --buildRequires Docker, Node.js 24, and an OpenAI-compatible model API key. Open http://localhost:3000 after startup. The admin panel is at /admin.

License and pricing

The code is licensed under MIT. The AIHOT name and logo are not covered by the MIT license. Fonts have their own licenses as noted in NOTICE. The project is open source and free to use.

這是什麼

AIHOT 是一個開源的網站框架,能夠自動從一批信源收集資料,使用大型語言模型進行篩選和評分,挑選出值得關注的內容,撰寫中文標題與摘要,將不同來源報導的同一事件聚類,計算熱點,並每天產出一份日報。它設計為可針對任何行業進行客製化,只需更換信源和精選標準。

主要功能

  • 支援六種信源:RSS、網頁列表、JSON 介面、X 帳號、微信公眾號,以及自訂腳本推送的內容。
  • 精選流程:預先篩選、兩次獨立評分(使用相同標準),並根據信源分級設定入選門檻。
  • 寫作:生成中文標題、答案先行的摘要、推薦理由、外文全文翻譯,並提取分類、標籤與事實。
  • 聚類:將不同來源報導的同一事件聚合成一個事件,後續進展掛在同一個事件下。
  • 熱點:根據獨立來源數量計算事件熱度,48 小時內衰減,並標示上升或新出現的主題。
  • 自動化報刊:每日、每週、每月自動編排日報、週報與月報。
  • 後台管理:信源管理、內容診斷、人工評分修正、精選評測、各步驟獨立模型設定、預算熔斷與操作記錄。

使用情境

  • 建立特定行業的熱點新聞網站(例如法律、人力資源、金融、貴金屬)。
  • 打造個人化的新聞聚合器,根據自訂標準篩選與摘要內容。
  • 為內部團隊或社群自動生成每日簡報。

快速開始

bashgit clone https://github.com/KKKKhazix/AIHOT.git myhotcd myhotnode scripts/init-env.ts --llm-keydocker compose up -d --build需要 Docker、Node.js 24 以及一個 OpenAI 相容的模型 API 金鑰。啟動後開啟 http://localhost:3000。後台位於 /admin。

授權與價格

程式碼採用 MIT 授權。AIHOT 的名稱與 Logo 不包含在 MIT 授權範圍內。字型有其各自的授權,詳見 NOTICE 檔案。此專案為開源且免費使用。

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/kkkkhazixaihot.svg)](https://yunodb.com/tools/kkkkhazixaihot?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/kkkkhazixaihot?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/kkkkhazixaihot.svg" alt="Featured on YunoDB" height="20"></a>