owencode — AI Coding Assistant TUI

A terminal UI coding assistant written in Rust, inspired by Claude Code. However it only uses around 13 MB of memory compared to around 330 MB for Claude code.
Supports Ollama (local) and Nvidia NIM (cloud) as AI backends.

Features

• Rich TUI — built with ratatui: chat history, file tree sidebar, status bar
• Dual backends — Ollama (local) and Nvidia NIM (OpenAI-compatible)
• Streaming — token-by-token streaming output with live display
• Tool use — shell execution, file read/write, directory listing
• File tree — sidebar shows current directory contents with icons
• Vim-like — Normal/Insert modes, j/k navigation
• Slash commands — /run, /read, /ls, /cd, /model, /clear
• Syntax highlighting — code blocks rendered with language labels
• Model switching — pop-up model picker for available models
• Persistent config — ~/.config/owencode/config.toml

Installation

Prerequisites

# Install Rust
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
# For Ollama backend (optional)
curl -fsSL https://ollama.ai/install.sh | sh
ollama pull llama3.2

Build & Run

git clone https://github.com/LoveIiei/Owen_Code.git
cd owencode
cargo build --release
./target/release/ocode

Or install globally:

cargo install --path . ocode

Configuration

Config is auto-generated at ~/.config/owencode/config.toml on first run:

default_backend = "ollama" # or "nim"
[ollama]
base_url = "http://localhost:11434"
default_model = "llama3.2"
[nim]
base_url = "https://integrate.api.nvidia.com/v1"
api_key = "nvapi-xxxxxxxxxxxx" # Get from build.nvidia.com
default_model = "meta/llama-3.1-70b-instruct"
[ui]
show_file_tree = true
syntax_highlight = true
mouse_enabled = true

Nvidia NIM Setup

1. Sign up at build.nvidia.com
2. Generate an API key
3. Set api_key in config or NVIDIA_API_KEY env var
4. Set default_backend = "nim" in config

Keybindings

Key	Action
i or a	Enter insert mode
Esc	Return to normal mode
↑↓ / j k	Scroll chat
g / G	Scroll to top / bottom
m	Open model selector
?	Help overlay
q / Ctrl+C	Quit

Slash Commands

Command	Description
/run <cmd>	Execute a shell command
/read <file>	Read a file into AI context
/ls [path]	List directory contents
/cd <path>	Change working directory
/model	Switch AI model (popup)
/clear	Clear chat history
/help	Show help
/quit	Exit

Architecture

src/
├── main.rs # Entry point + panic handler
├── app.rs # App state, event loop, command handling
├── config.rs # Config file management
├── events.rs # Event system (keyboard, AI stream, tick)
├── ai/
│   ├── mod.rs # AiBackend trait + shared types
│   ├── ollama.rs # Ollama streaming backend
│   └── nim.rs # Nvidia NIM (OpenAI-compat) backend
├── tools/
│   ├── mod.rs # Tool trait + ToolResult
│   ├── shell.rs # Shell command execution
│   └── file.rs # File read/write/list
└── ui/
    ├── mod.rs # Main draw() + layout
    ├── chat.rs # Chat history + input box
    ├── file_tree.rs # Sidebar file tree
    ├── status_bar.rs # Bottom status bar
    ├── model_select.rs # Model picker popup
    └── help.rs # Help overlay

Roadmap

• [ x ] Multi-line input editor
• Diff view for file edits
• Clipboard support
• [ x ] Command history (↑↓ in input)
• [ x ] Session save/restore
• [ x ] Config hot-reload
• Custom system prompt per session
• MCP (Model Context Protocol) tool support