essay / agents
Clawd.bot: this open-source project has 44,000 GitHub stars for a reason
Have not had this 'wait, this changes everything' feeling since ChatGPT dropped. It is what Siri was supposed to be 13 years ago.
Pulled an all-nighter I was not planning on. Have not had this “wait, this changes everything” feeling since ChatGPT dropped.
What is it
Clawdbot is an open-source, self-hosted AI assistant created by Peter Steinberger. You run it on your own machine (Mac Mini, laptop, Raspberry Pi, even a $5/month VPS) and interact with it through apps you already use: Telegram, WhatsApp, Discord, Slack, iMessage.
Why it is different
It is not sitting in a browser tab waiting for prompts. It runs 24/7, has persistent memory across conversations, and actually does things on your computer.
It handles OAuth flows on its own. Just tell it what service to connect and it figures it out. It can write its own code, give itself new skills, run tasks on a schedule, generate images, send voice messages, even call your phone. And it remembers everything you have talked about, so you pick up right where you left off.
The wildest part: it edits its own prompts and hot-reloads them. It literally teaches itself new tricks as you use it.
This is what Siri was supposed to be 13 years ago.
The responsible approach
You are giving an AI shell access. Start on a separate machine or cheap VPS, not your main laptop. Keep human-in-the-loop enabled. Give it read-only access to sensitive stuff until you trust it.
The power is real, so is the responsibility.
From an engineering perspective, the patterns in this codebase are worth studying regardless of whether you plan to run it. The architecture for persistent agent memory, multi-channel communication, and self-modifying behavior represents where personal AI assistants are heading. Understanding these patterns now gives you a head start on building or evaluating the next generation of agent infrastructure.