dgtlmoon / changedetection.io
The most popular self-hosted website change monitor, now supercharged with LLM-powered intent filtering and plain-language diff summaries that work with any model, including fully local ones.
星标趋势
AI 分析
项目摘要
changedetection.io is a popular self-hosted website change detection and monitoring platform that watches web pages for content changes and sends alerts via Discord, Email, Slack, Telegram, webhooks, and more. Recently it added AI-powered features built on LiteLLM, letting users connect any LLM (OpenAI, Gemini, Anthropic, Ollama, vLLM, LM Studio) to filter diffs with plain-English intent rules and generate plain-language change summaries. With 34k+ stars, 151 contributors, and frequent releases, it is one of the most battle-tested open-source monitoring tools available.
为什么值得关注
It is the de facto open-source standard for self-hosted website change monitoring, and its recent LLM integration is a pragmatic example of AI enhancing an existing tool: instead of raw diffs, users get intent-based filtering ('notify me only when the price drops below $50') and human-readable summaries, with full support for locally-run models via Ollama/vLLM for privacy-conscious users.
优势
- Battle-tested and actively maintained: 34k+ stars, 151 contributors, 15 releases in the last 6 months, CI pipeline, and first-class Docker support
- LLM-agnostic AI features via LiteLLM supporting 100+ providers, including fully local/self-hosted options (Ollama, vLLM, LM Studio, llama.cpp)
- Extensive notification integrations (Discord, Slack, Telegram, Email, Webhooks) and built-in Chrome browser for JavaScript-heavy pages
- Plain-English intent-based alert filtering dramatically reduces false positives and notification noise
- Generous free open-source core (Apache-2.0) with an optional low-cost SaaS plan
局限性
- AI capabilities are an enhancement layer rather than the core; the project is fundamentally a web monitoring tool, not an AI-native product
- AI features require either paid LLM API usage or the operational overhead of self-hosting a model
- Sending monitored page content to third-party LLM providers raises privacy considerations for sensitive watch targets
- Self-hosting still requires Docker familiarity and basic server administration
使用场景
- Price-drop monitoring with AI-filtered alerts like 'notify me only when the price falls below $50'
- Back-in-stock and restock alerts for e-commerce products
- Website defacement and unauthorized content change monitoring
- Receiving plain-language LLM summaries of changes instead of raw diffs (e.g., 'Price dropped from $89.99 to $67.00')