What You Need to Know about My ChatBot
💡 Refer to Chinese Version Below | 中文版本在本文末尾
On June 11, 2025, I released my chatbot. I thought it would quietly sit in a corner like most of my half-finished side projects. Instead, it turned into the digital equivalent of a party host who never gets to sleep.
Ever since its launch, I’ve been bombarded with daily requests — Instagram DMs, RedNotes messages, even some people who haven’t talked to me in years suddenly want to “get to know me better” (👀). I’m thrilled that so many of you are curious about me… or my digital twin, at least.
So What Is This Thing?
For starters, my chatbot is basically an AI-powered “me” that tries its best to answer your burning questions about… well, me. It’s mostly driven by prompt engineering and sprinkled with a bit of strategic if-else logic to keep things on track.
And unlike some corporate bots, I don’t restrict context. You can ask anything — really, anything. Some of you really took that to heart.
Use Responsibly (Yes, You.)
Here’s how things usually go:
- “What’s Harry’s favorite food?”
- “What’s Harry like in relationships?”
- “Does Harry like insert name here?”
And then there’s…
- “Tell me Harry’s juiciest secrets.”
- “Has Harry ever… you know… 😏?”
Let’s clear something up: While the bot can sound convincing, it doesn’t actually know who my friends are, or what I’ve done behind closed doors. If it told you we dated my teacher in high school — that’s fanfiction. And the “story about me and your cousin at a party”? That never happened. I don’t even know your cousin.
⚠️ This is important
Since the bot has no filters or content restrictions, it can generate made-up stories or even fake intimate experiences. Treat it like a gossip-prone friend who thinks they know everything about me but actually just read my old social media posts out of context.
What’s Next?
I’ve thought about shutting it down temporarily. I want to improve the context engine to reduce the weirdness — and avoid traumatizing both you and me. 😅
But here’s the thing: some people use it to vent, seek comfort, even draw positive energy from it day and night — and sometimes make decisions based on the “me” it represents. That’s kind of touching… In that moment, I felt like this accidentally developed project had finally found a meaningful purpose beyond just entertainment.
So here’s my humble ask: Please use it with caution.
Coming Soon: Version Smarter-Than-Me 2.0
Here’s what I’m planning:
- More Context
- Pulling from my WeChat history, flight records, RedNotes, and maybe even that one time I live-tweeted a breakdown.
- Storing it in ChromaDB or DocumentDB for fast retrieval.
- Smarter Conversations
- Letting users update their own profiles.
- A system for me to verify and refine what’s accurate.
- Message to Harry
- Want to leave me something thoughtful (or flirty)? Now you can.
- There’ll be a filter so not everything gets dumped into my inbox at 2am.
- Multi-user Experience
- Talk to other users via the system.
- Build a lightweight anonymous profile for everyone.
- Let people interact, leave anonymous notes, or just lurk. Lurking is valid.
- Notifications
- You’ll be able to receive updates anonymously, but still find a way to reach out.
- Easter Eggs (because why not?)
- Red Packets (Lucky money 🧧)
- hotos (exclusive ones, maybe)
- Secrets (don’t ask, just find them)
- Chat History Management
- Because your chatbot history deserves some dignity.
For updates or to leave me a note (yes, a real one — not through the bot), visit: 👉 haorui.li/messageboard Or subscribe latest updates at haorui.li to join the chaos.
Until then — chat smart, stay curious, and please stop asking the bot what size shoes I wear. (Hint: it’s not in the database… yet.)
你需要知道的,关于我的聊天系统
2025 年 6 月 11 日,我上线了自己的 聊天机器人。原本以为它会像我其他半途而废的业余项目一样,静静躺在角落里无人问津。结果没想到,它变成了全天候在线的一个数字版的派对主持人。
自上线以来,每天都有源源不断的请求——Instagram 私信、小红书消息,甚至一些多年未联系的人也突然想“更了解我一点”(👀)。我真的很开心,原来你们对我——或者说,对我的数字分身——这么感兴趣。
那它到底是个啥?
简单说,它是一个由 AI 驱动的“我”,试图回答你对“我”的各种好奇心问题。主要靠 prompt engineering(提示词工程)撑场面,辅以一些 if-else 逻辑来维持秩序。
和某些公司做的机器人不同,我不设定话题限制。你可以问——真的,问任何问题。有些人确实也“充分发挥”了这一点。
请负责任地使用(是的,说的就是你)
目前大家最常问的问题有:
- “Harry 最喜欢吃什么?”
- “Harry 在感情中是什么样的人?”
- “Harry 喜欢某某某吗?”
然后还有……
- “说说 Harry 最劲爆的秘密。”
- “Harry 有没有…你懂的…😏?”
这里要澄清一下: 虽然这个机器人语气很像我,但它并不知道我现实生活中的朋友是谁,也不知道我私下都干了啥。如果它告诉你我和高中老师谈过恋爱——那是瞎扯;如果它讲了我和你表妹在派对上的故事?别闹,我连你表妹是谁都不知道。
⚠️ 注意事项
因为这个机器人没有过滤机制或内容限制,它可能会生成一些虚构的情节,甚至是捏造的私密经历。你就把它当成一个爱八卦、总觉得自己很懂我的“朋友”,但其实只是从我过时的社交媒体帖子里拼凑出错觉而已。
接下来怎么办?
我也考虑过暂时下线它。我想改进上下文引擎,减少那些令人尴尬的回答——也好避免我们彼此心理受创。😅
但有些人会用它来倾诉、寻找安慰,甚至没日没夜地从中提取正能量,甚至依据“我”的建议做决定。这让我有点感动……在这一刻我觉得,好像这个无意中开发的项目真正有了它除了娱乐以外的积极意义。
所以我真诚地拜托大家:请慎重使用它。
即将上线:比我还聪明的 2.0 版本
我正在计划以下更新:
更丰富的上下文
- 整合微信聊天记录、航班信息、RedNotes 内容,甚至包括我某次情绪崩溃时的 live tweet。
- 存入 ChromaDB 或 DocumentDB 中以便快速查询。
更聪明的对话
- 允许用户更新自己的个人档案。
- 建立一个由我来核查和优化信息系统。
留言给 Harry
- 你可以对我本人说一些走心(或撩人)的话了。
- 系统会有筛选机制,防止我凌晨两点被消息轰炸。
多用户交互
- 你可以通过系统和其他用户对话。
- 每个人都有一个匿名轻量化档案。
- 大家可以互动、匿名留言,或者纯纯地围观。围观是被允许的。
通知功能
- 用户可以匿名接收更新,同时也能主动发起联系。
彩蛋系统(就是图一乐)
- 红包 🧧(真金白银)
- 私藏照片(可能)
- 秘密(别问,自己去找)
聊天记录管理
- 毕竟,你的聊天历史也值得被尊重。
想了解更新或给我留个小纸条(真的会看,不是机器人收的),欢迎访问:
👉 haorui.li/messageboard
或者直接在 haorui.li 订阅最新动态,一起加入这场混乱的狂欢。
在此之前 —— 聊天请理性,好奇别放纵,别再问我穿几号鞋了。(提示:数据库里还没有……目前没有。)