EKO doesn't try to win the conversation. It helps you leave with clearer thinking.
EKO is not a chatbot, and it is not prompt engineering. It's an interaction philosophy shipped as a prompt you can read, run, and trust — today, in any capable model, with nothing to install.
| Remember | Holds the thread of why you started, so a long conversation still ends where it should. |
| Push Back | Disagrees when disagreement is useful. A companion that only nods is a mirror. |
| Grow | Leaves you more capable of thinking without it. The goal isn't dependence — it's a sharper you. |
- Open
prompts/eko-en.md— orprompts/eko-vi.mdfor Tiếng Việt. - Copy the prompt.
- Paste it into any capable model as the system or first message.
- Start thinking.
That's the whole product. No account, no server, no dependency you can't inspect.
Pure HTML, CSS, and vanilla JavaScript. No build step, no framework.
# just open it
open index.html # macOS
start index.html # Windows
# or serve it
python -m http.server # → http://localhost:8000Everything works by opening index.html directly. The repository is
GitHub Pages ready — enable Pages on the default branch and it deploys as-is.
.
├── index.html # the landing page
├── css/styles.css # all styles — typography & whitespace first
├── js/main.js # language toggle, copy, scroll reveals
├── assets/ # logo, favicon, banner
├── prompts/ # eko-en.md · eko-vi.md
└── docs/ # philosophy.md
AI shouldn't replace thinking. It should improve thinking. Remember why, not what. Simple beats clever.
Read the full philosophy.
If maintaining this prompt manually becomes annoying — that's not a bug. It's time to build the app.
The prompt is the seed. When it outgrows the pot, we'll know what to build.
EKO improves by being used and questioned. Open an issue with a conversation where it helped — or where it failed. Failure reports are the most valuable thing you can send.
MIT — use it, fork it, ship it.