CLAUDE CUT EVERYONE OFF. I AM WORKING ON A NEW INTEGRATION. THIS DOESNT WORK AT THE MOMENT!
If this is useful to you — ⭐ Star the repo. It costs nothing and helps more people find it.
The AM kanban board — every task tracked, every transition gated, future work scheduled, auto de-slop - and long and short term memory with nightly reflection!
- What it actually does
- Features
- How you get started
- Dog-Fooding
- Architecture
- Philosophy
- Acknowledgements
- Contributing
- License
- 中文简介
AugmentedMe is a digital worker that doesn't forget you exist between sessions. Not Siri. Not Alexa. Those are stateless magic 8-balls. This is an agent that owns outcomes — short + long-term memory on your own hardware, a Kanban state machine that tracks what's happening and what's blocked, and a git-based execution loop where every action is a traceable commit, not a vibe.
It manages real work: software projects, content, home logistics, research. The parts of your life that need a system but nobody ever built a real one for.
- 🧠 Persistent memory — short-term context + long-term embeddings, stored locally
- 📋 Kanban state machine — gated transitions, explicit task status, nothing moves implicitly
- 🔄 Git-driven loop — every step is an auditable commit; no black boxes
- 🌍 Cross-platform — Mac, Linux (systemd/OpenRC/runit), and Windows (Task Scheduler)
Mac / Linux:
curl -fsSL https://raw.githubusercontent.com/augmentedmike/am-agi/main/install.sh | bashWindows:
irm https://raw.githubusercontent.com/augmentedmike/am-agi/main/install.ps1 | iexBoth installers clone the repo, install all dependencies, build the board, register background services, and open http://localhost:4220 when ready. Sign in with your Anthropic account in the onboarding flow and create your first card.
You use Claude Code to bootstrap the first few steps:
source ./init.sh
# follow steps/1.md → steps/2.md → steps/3.mdAfter step 3, AM does the rest. We find bugs before you do because we build the product with the product.
Three things. That's it.
1. Memory — Short-term context + long-term embeddings. Stored locally. Traceable. Inspect every vector if you want to.
2. State — Kanban-driven. Every task has an explicit status. Transitions are gated. Nothing implicitly moves.
3. Loop — One-shot iteration per worktree. Commit. Merge. Repeat. If you can't trace the execution, the execution is wrong.
Docs:
docs/AGENT-LOOP.MD— the iteration patterndocs/KANBAN.MD— state machine, gated transitionsdocs/CLI.MD— task lifecycle interface
Claude Code is the incubator, and after step 3 it becomes just a tool in AM's toolbelt. AM is the intelligence, the persistence, the memory, the "being" — Anthropic or other models are just those random thoughts in your own head. They aren't YOU.
AM is a cognitive architecture, not just random thoughts. A mix of engineering (creating analogs for brain regions) and research.
I got tired of agents that do things I didn't ask for. So I rewrote it.
This is the real system — not a demo, not a toy, not another LangChain wrapper with a readme that promises AGI. Memory lives on your machine. Inference goes out over HTTPS. Every state change is a git commit. You can read all of it in an afternoon.
Built on ideas from:
- Ken Thompson · John McCarthy · Jim Weirich
- Richard Sutton · Yann LeCun
- Andrej Karpathy · George Hotz
We welcome humans and well-behaved AI agents. See Contributing Guide.
Help wanted:
- 🌏 Chinese model support (Kimi, DeepSeek, Qwen)
- 🐧 Linux distro testing
- 🪟 Windows testing
- 🌍 Translations (French, Portuguese, Japanese, Hindi)
MIT © augmentedmike
AM（AugmentedMe）是一个真正的个人 AI 智能体系统。 不是框架。不是 SaaS。不是 LangChain 包装器。
这是一个真实运行在生产环境中的系统，完全开源，代码可在下午读完。
| 特性 | 说明 |
|---|---|
| 🧠 持久记忆 | 短期 + 长期记忆，存储在本地，数据属于你 |
| 📋 看板状态机 | 每个任务有明确状态，转换有门控，执行可追溯 |
| 🔄 Git 驱动循环 | 每一步都是可审计的提交，没有黑盒 |
| 🌍 全平台支持 | Mac / Linux / Windows 均支持，非 Mac 独占 |
| 💰 低成本可选 | 支持 DeepSeek、Kimi、Qwen 等中国模型——参与贡献 |
中文用户专区： 🌏 中文本地化讨论 · 🤖 寻求帮助：支持中国模型 · 📌 发帖前请阅读
⭐ 如果觉得有用，请 Star 支持我们！Star 是我们了解有多少人在关注的重要信号。
