ChaTTY is a Terminal User Interface (TUI) for chatting with AI models (OpenAI, Gemini), written in Rust. It brings the power of ChatGPT and Gemini to your terminal, with features like conversation management, multiple model support, and intelligent context compression.
- Multiple AI Models: Support for OpenAI and Gemini
- Conversation Management: Save and restore chat histories
- Smart Context Compression: Automatically manages long conversations (experimental)
- Syntax Highlighting: Beautiful code block colorization
- Fast and Efficient: Written in Rust for optimal performance
- Copy most recent message
Ctrl + e >> c - Press
yto copy selected message.
# Download from GitHub https://github.com/x00real/chatty/releases
# or
# Install from Crate.io
$ cargo install --locked chatty-rs
# Build from source
$ git clone https://github.com/x00real/chatty && cd chatty
# The output will be in ./builds/usr/bin/chatty
$ make DESTDIR=builds installChaTTY requires minimal configuration before first use. You'll need to provide at least one backend connection.
Create a config file in one of these locations:
$XDG_CONFIG_HOME/chatty/config.toml$HOME/.config/chatty/config.toml$HOME/.chatty.toml
[[backend.connections]]
enabled = true
alias = "OpenAI"
kind = "openai"
endpoint = "https://api.openai.com"
api_key = "<your_api_key>"View the complete configuration options in our default config file.
$ chatty --help
A Terminal UI to interact OpenAI models
Default configuration file location looks up in the following order:
* $XDG_CONFIG_HOME/chatty/config.toml
* $HOME/.config/chatty/config.toml
* $HOME/.chatty.toml
Usage: chatty [OPTIONS]
Options:
-c, --config <PATH>
Configuration file path
-v, --version
Show the version
-h, --help
Print help (see a summary with '-h')Contributions are welcome! Feel free to:
- Report bugs
- Suggest new features
- Submit pull requests
This project is licensed under the MIT License - see the LICENSE file for details.
- Built with ratatui
- Inspired by ChatGPT and other terminal-based tools