t1code Terminal UI Skill

---
name: t1code-terminal-ui
description: AI-powered terminal coding assistant (T3Code in your terminal) using OpenTUI
triggers:
  - "use t1code in my terminal"
  - "run t1code TUI"
  - "set up terminal AI coding assistant"
  - "t1code configuration"
  - "bunx t1code"
  - "T3Code terminal version"
  - "t1code LLM coding TUI"
  - "install t1code globally"
---

# t1code Terminal UI Skill

> Skill by [ara.so](https://ara.so) — Daily 2026 Skills collection.

## What is t1code?

t1code is a terminal user interface (TUI) for AI-assisted coding, inspired by T3Code (by @t3dotgg and @juliusmarminge). It brings an LLM-powered coding assistant directly into your terminal using the OpenTUI framework. It supports models via API (similar to Codex/Claude/OpenAI-compatible endpoints) and runs entirely in the terminal.

---

## Installation

### Run instantly (no install)

```bash
bunx @maria_rcks/t1code

Install globally

bun add -g @maria_rcks/t1code

After global install, run with:

t1code

Develop from source

git clone https://github.com/maria-rcks/t1code.git
cd t1code
bun install
bun dev:tui

---

Requirements

• Bun runtime (v1.x+)
• A terminal that supports ANSI escape codes
• An LLM API key (OpenAI-compatible endpoint recommended)

---

Key Commands

Command	Description
bunx @maria_rcks/t1code	Run t1code without installing
bun add -g @maria_rcks/t1code	Install globally
t1code	Launch TUI (if globally installed)
bun dev:tui	Run in dev mode from source
bun install	Install dependencies from source

---

Configuration

t1code uses environment variables for API keys and model configuration. Set these before running:

# OpenAI-compatible API key
export OPENAI_API_KEY=your_api_key_here

# Optional: custom base URL for OpenAI-compatible APIs (e.g. local Ollama, Together, etc.)
export OPENAI_BASE_URL=https://api.openai.com/v1

# Optional: specify default model
export OPENAI_MODEL=gpt-4o

You can place these in a .env file in the project root when developing from source:

OPENAI_API_KEY=your_api_key_here
OPENAI_BASE_URL=https://api.openai.com/v1
OPENAI_MODEL=gpt-4o

---

TUI Navigation

Once launched, t1code presents a terminal UI with:

• Chat/prompt input — Type your coding question or instruction
• Code output panel — Displays AI-generated code with syntax highlighting
• Keyboard shortcuts — Navigate panels, copy output, clear history

Common TUI interactions:

Key	Action
Enter	Submit prompt
Ctrl+C	Exit t1code
Tab	Switch focus between panels
Ctrl+L	Clear the chat/output
Arrow keys	Scroll through output

---

Development: Project Structure

t1code/
├── src/
│   ├── tui/          # OpenTUI components and layout
│   ├── llm/          # LLM API client logic
│   ├── config/       # Configuration loading
│   └── index.ts      # Entry point
├── assets/
│   └── repo/
├── package.json
└── bun.lockb

---

Real Code Examples

Running t1code programmatically (from source entry point)

// src/index.ts - typical entry pattern
import { startTUI } from "./tui";
import { loadConfig } from "./config";

const config = loadConfig();
await startTUI(config);

Extending the LLM client (OpenAI-compatible)

// src/llm/client.ts - example OpenAI-compatible fetch
const response = await fetch(`${process.env.OPENAI_BASE_URL}/chat/completions`, {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    Authorization: `Bearer ${process.env.OPENAI_API_KEY}`,
  },
  body: JSON.stringify({
    model: process.env.OPENAI_MODEL ?? "gpt-4o",
    messages: [
      { role: "system", content: "You are a helpful coding assistant." },
      { role: "user", content: userPrompt },
    ],
    stream: true,
  }),
});

Using a local Ollama model with t1code

# Start Ollama with a code model
ollama run codellama

# Point t1code to local Ollama
export OPENAI_BASE_URL=http://localhost:11434/v1
export OPENAI_API_KEY=ollama
export OPENAI_MODEL=codellama
bunx @maria_rcks/t1code

Using Together AI or other OpenAI-compatible providers

export OPENAI_API_KEY=$TOGETHER_API_KEY
export OPENAI_BASE_URL=https://api.together.xyz/v1
export OPENAI_MODEL=codellama/CodeLlama-34b-Instruct-hf
bunx @maria_rcks/t1code

---

Common Patterns

1. One-off code generation session

OPENAI_API_KEY=your_key bunx @maria_rcks/t1code

2. Persistent global install with config in shell profile

# Add to ~/.bashrc or ~/.zshrc
export OPENAI_API_KEY=your_key
export OPENAI_MODEL=gpt-4o

# Then just run
t1code

3. Contributing a new TUI component (from source)

// src/tui/MyPanel.ts - OpenTUI component pattern
import { Box, Text } from "opentui";

export function MyPanel({ content }: { content: string }) {
  return (
    <Box border="single" padding={1}>
      <Text>{content}</Text>
    </Box>
  );
}

4. Adding a new keyboard shortcut (from source)

// src/tui/keybindings.ts
import { onKey } from "opentui";

onKey("ctrl+r", () => {
  // Reload/reset logic
  resetChat();
});

---

Troubleshooting

bunx not found

Install Bun: https://bun.sh

curl -fsSL https://bun.sh/install | bash

TUI renders incorrectly / garbled output

• Ensure your terminal supports 256-color ANSI (iTerm2, Warp, kitty, Windows Terminal, etc.)
• Try resizing your terminal window — some TUI layouts require minimum dimensions
• On Windows, use Windows Terminal or WSL2

API errors / no response

• Verify OPENAI_API_KEY is set correctly
• Check that OPENAI_BASE_URL points to a valid OpenAI-compatible endpoint
• Test your API key independently:

curl https://api.openai.com/v1/models \
  -H "Authorization: Bearer $OPENAI_API_KEY"

bun dev:tui fails with module errors

# Clean install
rm -rf node_modules bun.lockb
bun install
bun dev:tui

Slow streaming / no streaming output

• Confirm your provider supports SSE streaming on the /chat/completions endpoint
• Some self-hosted models (e.g., older Ollama builds) may need stream: false in the request body

---

Links

• npm: https://www.npmjs.com/package/@maria_rcks/t1code
• GitHub: https://github.com/maria-rcks/t1code
• OpenTUI: https://github.com/nickvdyck/opentui
• Original T3Code: https://github.com/t3dotgg
• License: MIT