📄 中文摘要
ContextMD 是一个 Python 中间件，旨在为 OpenAI、Anthropic 和 LiteLLM API 调用添加持久记忆。通过将对话存储为人类可读的 Markdown 文件，ContextMD 能够自动提取事实并将其重新引入未来的请求中。这一技术解决了大语言模型（LLM）在对话中缺乏记忆的问题，使得应用程序能够更好地理解和记住用户的历史交互，从而提升用户体验和交互的连贯性。开发者可以利用 ContextMD 来增强其 LLM 应用的智能性和实用性。
永不重复自己：通过 ContextMD 为 LLM 应用程序提供持久记忆
📄 English Summary
Never Repeat Yourself: Give Your LLM Apps Persistent Memory with ContextMD
ContextMD is a Python middleware designed to add persistent memory to OpenAI, Anthropic, and LiteLLM API calls. By storing conversations in human-readable Markdown files, ContextMD can automatically extract facts and bootstrap them back into future requests. This technology addresses the memory limitation of large language models (LLMs) in conversations, enabling applications to better understand and remember users' historical interactions, thereby enhancing user experience and the coherence of interactions. Developers can leverage ContextMD to improve the intelligence and practicality of their LLM applications.