KKKKhazix/AIHOT

What is it?

Key features

• 01Collects from RSS, web pages, JSON APIs, X accounts, WeChat accounts and pushed content.
• 02Pre-filters items and scores them twice with an LLM before selection.
• 03Clusters reports on the same event and ranks trends by distinct sources.
• 04Generates Chinese titles, summaries and daily, weekly and monthly digests.
• 05Exposes RSS, a public API, MCP and llms.txt for agents.
• 06Prompts and selection thresholds are editable plain files.

Use cases

• Running an industry news site for law, HR, finance or precious metals.
• Building a private daily briefing that tracks specific sources.
• Experimenting with LLM-based content curation and event clustering.

Quick start

1. 1

   Clone the repository

   git clone https://github.com/KKKKhazix/AIHOT.git myhot
2. 2

   Enter the project folder

   cd myhot
3. 3

   Generate the config with your LLM API key

   node scripts/init-env.ts --llm-key <你的模型 API Key>
4. 4

   Start the site with Docker Compose, then open http://localhost:3000

   docker compose up -d --build

Things to know

• !Needs an OpenAI-compatible LLM API key, and Node.js 24 is required to generate the config.
• !It is the engine of the author's AIHOT site rather than a polished general-purpose framework, and it ships with only 18 demo sources.
• !The author says they are not a professional developer, so rough edges are likely.

Featured in Repolode Weekly

• Issue #3 · Big MoE models streamed from SSD, and millisecond Rust rebuilds

  A self-hosted framework for an industry news site pulling from RSS, web pages, APIs, X and WeChat. Items are pre-filtered, scored twice by an LLM, and clustered by event, with trends ranked by distinct sources.

In collections

Topics

FAQ

What is AIHOT?

A self-hosted framework for building an industry news site: you plug in your own sources and curation criteria, and it collects, filters, clusters and publishes daily digests. AIHOT is an engine for a curated news site. It pulls material from RSS feeds, web lists, JSON APIs, X accounts, WeChat public accounts and custom pushed content, then uses an LLM to pre-filter items, score them twice independently, and write Chinese titles and summaries. Reports from different outlets about the same event are grouped into one event, and trending is ranked by the number of distinct sources covering it rather than by article count.

Is AIHOT free to use? What license does it have?

AIHOT is open source under the MIT license. Check the license terms before using it in commercial products.

Is AIHOT still maintained?

Actively maintained. The last commit to AIHOT was on Nov 5, 2026.

What can I use AIHOT for?

Running an industry news site for law, HR, finance or precious metals. Building a private daily briefing that tracks specific sources. Experimenting with LLM-based content curation and event clustering.

How do I install AIHOT?

Clone the repository: git clone https://github.com/KKKKhazix/AIHOT.git myhot Enter the project folder: cd myhot Generate the config with your LLM API key: node scripts/init-env.ts --llm-key <你的模型 API Key> Start the site with Docker Compose, then open http://localhost:3000: docker compose up -d --build

Where can I get AIHOT?

The source code and installation instructions are on GitHub: https://github.com/KKKKhazix/AIHOT