A personal AI system that learns who you are, remembers what you've done, and gets smarter every time you use it.
Most people use AI like a search engine. Start fresh every time, re-explain everything, get generic answers.
An AI system changes that. It gives any AI tool long-term memory, personalized context, and compound intelligence over time.
You're already using an AI agent. ChatGPT is one. Claude Code is one. Gemini is one. This repo is what every agent should run on: the infrastructure that makes them know you, remember what you've done, and get better the more you use them.
This repo is the open-source companion to my YouTube channel, where I build this system in public and teach you how to build your own.
Two ways in. Pick yours.
Grab the Starter Kit. Paste one prompt into ChatGPT, Claude, or Gemini. It walks you through building a lightweight version of this system in any AI. No install required.
git clone https://github.com/1269/ai-system.git
cd ai-system
./install.shYou now have a working AI system on your machine. Fork it, rip it apart, make it yours.
Nine components, three tiers.
The irreducible core. Without these, there is no system.
| # | Component | What It Is |
|---|---|---|
| 1 | Input & Capture | How you talk to your agent. Typing, voice, transcripts, uploads. The surfaces where you show up (CLI, Slack, chat). |
| 2 | Context Engineering | Your agent's operating manual. The rules, skills, and system prompts that shape behavior. |
| 3 | Memory | Your agent's notebook. Facts about you, preferences, and pointers into knowledge, accumulated over time. |
| 4 | The Model | Your agent's brain. Claude, GPT, Gemini, or a local model. Interchangeable. |
What the agent can actually do beyond talking.
| # | Component | What It Is |
|---|---|---|
| 5 | Skills | Reusable capabilities. Packaged workflows your agent can invoke. |
| 6 | Integrations | Tools your agent can reach. Calendar, tasks, email, Slack, GitHub. |
| 7 | Knowledge Base | The library your agent references. Your notes, research, and frameworks, organized for retrieval. |
What makes a system compound over time.
| # | Component | What It Is |
|---|---|---|
| 8 | Orchestration | Multiple agents coordinating on a task. |
| 9 | Automation | Agents running on schedules, without you asking. |
Agents are not components. They're runtimes. ChatGPT, Claude Code, and the custom agents you build all use the nine-component system. Swap the agent, keep the system. Your memory, your knowledge, your skills stay yours.
The repo mirrors the framework. Each numbered folder is a component.
ai-system/
├── 01-capture/ Input & Capture
├── 02-context/ Context Engineering
├── 03-memory/ Memory
├── 04-model/ The Model
├── 05-skills/ Skills
├── 06-integrations/ Integrations
├── 07-knowledge/ Knowledge Base
├── 08-orchestration/ Orchestration
├── 09-automation/ Automation
│
├── starter-kit/ 15-min no-code entry
├── templates/ Copy-paste files for your own system
├── guides/ Deep dives, one per component
└── examples/ Real-world skills, integrations, setups
Each guide maps to a component and ships alongside a video.
| # | Component | Guide | Video |
|---|---|---|---|
| — | Overview | The 9 Components | Coming soon |
| 1 | Input & Capture | Guide | — |
| 2 | Context Engineering | Guide | — |
| 3 | Memory | Guide | — |
| 4 | The Model | Guide | — |
| 5 | Skills | Guide | — |
| 6 | Integrations | Guide | — |
| 7 | Knowledge Base | Guide | — |
| 8 | Orchestration | Guide | — |
| 9 | Automation | Guide | — |
- Knowledge workers who use AI daily and want it to actually know them
- Builders who want a system, not scattered conversations
- Teams who want to give their AI tools shared memory and real context
- Curious humans who think "there has to be a better way to use this stuff"
MIT. Use it, fork it, make it yours. If you ship something cool, tell me.