AgentCloak 发布，可在浏览器内使用 ChatGPT、Claude 及中国 AI 服务而不泄露真实数据。它在提示词发出前将敏感值替换为逼真的替身，收到响应后在本地换回原值，完全在浏览器内运行，现已免费使用，网址 http://agentcloak.ai/。作者指出 AI 隐私的核心是语境保留问题，单纯删减提示词会改变推理任务，难点是在去除身份信息的同时保留足够语义让模型正常作答。
People underestimate how much AI privacy is really a context-preservation problem.
Anyone can protect a prompt by deleting half of it.
the hard part is removing identity without changing the reasoning task.
AgentCloak launched to solve that problem in the browser.
It sits in the browser, replaces sensitive values with realistic stand-ins before the prompt leaves your machine, then puts the original values back into the response locally.
This isn't just redaction. Redaction removes information and changes the task. The harder problem is preserving enough semantics for the model to answer properly while removing the pieces that identify you.
Introducing AgentCloak: Use any AI without sharing your real data. Chinese AI services, ChatGPT, Claude, doesn't matter. You probably try to hide details before asking: different names, fake numbers, no address. But then the answer's useless because the AI is missing actual context. AgentCloak runs in your browser. It swaps your sensitive info for realistic fakes before sending anything, then swaps your real info back into the response. You get what you need. The AI gets nothing about you. Works entirely in-browser. Already trusted by some of the biggest companies in the world. Now free to use. http://agentcloak.ai/在 X 查看被引用的帖子
来源：Rohan Paul · x.com