KKKKhazix/AIHOT

Project overview

Because all selection prompts and entry thresholds are public and editable without code changes, the project makes the otherwise industry-specific judgment of which sources matter and what counts as hot explicit and customizable.

Project type

Workflow

Deployment

Refer to project documentation

License

MIT

Best for

• Developers who want to run their own industry-specific Chinese-language news and trending-topics site and are willing to edit prompts and configuration to define relevance and hotness.
• Teams building agent workflows that need agent-readable news content via RSS, API, MCP, Agent Markdown, or llms.txt.

Key capabilities

• Pre-screens material, scores twice with the same standard via prompts, and writes Chinese titles and summaries with translation; all prompts and thresholds are editable without code changes.
• Exposes RSS feeds, a public API, MCP, Agent Markdown, and llms.txt so agents can consume the same content.

Limitations and risks

• Operation requires an OpenAI-compatible model API key (DeepSeek, Qwen, and Zhipu all work); data leaves the local deployment for inference.
• Using the project requires coding ability.

Getting started

• Setup requires Docker, Node.js 24, and a model API key, followed by a four-command CLI setup.

Evidence and sources

• GitHub project description
• README
• README
• README
• README

Source updated at: 2026-10-15T20:12:58Z

Open original GitHub project