我为任何大型语言模型构建了一个本地记忆层——存储你的偏好,并在每个会话中注入它们
📄 中文摘要
每次与 AI 进行对话时,用户都需要重新解释自己的偏好,例如编程语言的选择、代码风格和认证方式等。这种重复的过程不仅耗时,而且效率低下。为了解决这个问题,开发者创建了一个名为 recall 的工具,允许用户存储个人偏好,并在每次会话中自动注入这些信息。用户只需通过简单的命令将偏好记录下来,随后可以方便地将其粘贴到任何 AI 聊天中,从而提升交互的个性化和效率。
📄 English Summary
I built a local memory layer for any LLM — stores your preferences, injects them into every session
Every interaction with AI starts from scratch, requiring users to repeatedly explain their preferences, such as programming language choices, coding styles, and authentication methods. This repetitive process is time-consuming and inefficient. To address this issue, a tool called recall has been developed, allowing users to store their preferences and automatically inject them into every session. Users can easily record their preferences using simple commands and conveniently paste them into any AI chat, enhancing personalization and efficiency in interactions.
Powered by Cloudflare Workers + Payload CMS + Claude 3.5
数据源: OpenAI, Google AI, DeepMind, AWS ML Blog, HuggingFace 等