Introduction
YECI (Your Engine for Coding Intelligence) is a Docker image that gives you a ready-made terminal workspace for AI-assisted coding. Start it, open a shell, and the major AI coding agents, several language runtimes and a complete terminal toolchain are already installed and configured.
Who it is for
- You use AI coding agents such as Claude Code, Codex, Gemini CLI or Grok, and want them to run in an isolated environment instead of directly on your machine.
- You want the same tools and versions every time, on every machine.
- You work in the terminal: tmux, Neovim, git.
What's inside
Current image: 20260928, updated 2026-09-28.
| Component | Version |
|---|---|
| Swift | 6.4.0 |
| Objective-C | gobjc (GCC) + GNUstep Base |
| Node.js | 26.10.0 |
| Go | 1.27.1 |
| Python | 3 + pip |
| Claude Code | 2.1.283 |
| Codex | 0.158.0 |
| Gemini CLI | 0.61.0 |
| Grok | 1.0.41 |
| Neovim | 0.12.5 (with the yeci.nvim configuration) |
| yeci-notify | 1.1.3 |
| tmux, bashtidy | latest at build time |
The base system is Ubuntu 24.04 (the official Swift image). Common command-line tools are included too: git, curl, jq, ripgrep (rg), make, tree, zip, nano, and network tools (ping, dig, ss, nc).
Design choices
- Pinned versions. Every tool version is written into the image definition, so a given image tag always contains exactly the same tools.
- Updated automatically. New versions of the tools are checked every day. When something changes, a new image is built and published with a dated tag. See Images & Tags.
- Terminal only. There is no desktop environment and no Xcode. The container is where you write and run code with AI agents; your own machine remains the place for GUI tools.
- Your code stays on your machine. Your project directory is mounted into the container at
/workspace, so files are edited in place and survive when the container is removed.
Next step
Go to Quick Start to run the container.