AI UNDERDOGSDAILY PICK
AI UNDERDOGS
AI 每次聊完就失忆
AI forgets everything
Project Brain
你有没有发现
用 Claude Code 写了一下午
新开一个对话，它什么都不记得了
You spend three hours pairing with Claude Code
Open a new chat — it's like you
never existed
给 AI 加一层持久记忆
A persistent memory layer
Project Brain 在你的编辑器外面搭了一个记忆层
对话结束了，记忆还在
下次打开自动加载
Project Brain attaches a memory layer to your
AI editor. Sessions end, but the context survives
Next time you open a chat, it picks
up where you left off
★ SIGNAL 1
记忆不在对话里
Memory lives outside the chat
关键设计：记忆存在对话之外
不是依赖聊天记录
独立维护一份持久化上下文
新对话启动时自动注入
The memory isn't stored inside any chat thread
It's a standalone context that gets injected into
every new session — so closing a window
doesn't erase what your AI learned
★ SIGNAL 2
一个记忆，三个编辑器
One memory, three editors
同一份记忆层同时服务 Claude Code
Windsurf 和 Cursor
换工具不用重头来，项目上下文跟着你走
One memory layer serves Claude Code, Windsurf, and
Cursor all at once. Switch editors mid-project —
your context follows you instead of dying with
the last chat
项目还在早期，没挂什么营销话
但从 v2ex 到 HN
每个用 AI 写代码的人都在骂同一个事——上下文一关就丢
这个需求是真痛
It's early — no landing page, no pitch
deck. But every AI coder on v2ex and
HN complains about the same thing: context vanishes
the moment you close the window. This project
tackles that head-on
AI UNDERDOGS
新对话一开，上下午的上下文全没了
New chat, zero context — every single time
Project Brain
关注 · 每天发现更多 AI 神作
raw.githubusercontent.com/OoneBreath/claude-code-project-brain/main/docs/demo.gif