ht_demo_2.3.4.mp4
ht is a shell helper tool focused on simplicity that can generate, explain and execute shell commands using AI.
-
ht <instruction> - answers with shell command
-
ht e|explain - explains last answer
-
ht e|explain [command] - explains command
-
ht x|execute - executes last answer
ht find all IPv4 addresses in file A and write to file Bht explainht explain ls -lSht explain "ps -aux | grep nvidia"ht execute
- Low token usage
- Cached responses
- Easy installation (no root required)
- Automatic updates
I initially created ht as a simple experiment to test GPT3's usefulness with shell commands. However, I now find myself using it extensively in my daily tasks. So I'm sharing it with the hope that it can benefit others in the same way. It's using OpenAI's GPT3.5-Turbo model now and I plan to add more models in the future, including locally running models.
ht is written in Dart. As a result, it is compiled into a single, self-contained binary. This means that the ht binary operates independently without requiring any external dependencies or runtime environments. Linux (x86_64) and MacOS (arm64 & Intel) binaries are available for download.
To use ht, you'll need an OpenAI API key. The good news is that due to ht's low token usage, a typical request costs about $0.00025, making it an incredibly budget-friendly tool for daily usage. You can # for an API key here or refer to this article for detailed instructions.
ht communicates directly with OpenAI's API, without involving a third-party server. For automated updates to work, ht will send a request to the GitHub API to check for new releases.
- Download the archive for your platform from the Downloads section below.
- Unzip the archive.
- Using a terminal, navigate to the directory containing the ht binary and run it with the -i flag to start the installation process.
cd Downloads
./ht_2-0-3_linux64 -i
ht will be installed to '~/.config/ht' and the directory will be added to your PATH. Future updates will be installed automatically.