wut

CLI that explains the output of your last command.

Just type wut and an LLM will help you understand whatever's in your terminal. You'll be surprised how useful this can be. It can help you:

• Understand stack traces
• Decipher error codes
• Fix incorrect commands
• Summarize logs

Installation

> pipx install wut-cli

Once installed, you can use OpenAI or Claude as your LLM provider. Just add the appropriate API key to your environment:

> export OPENAI_API_KEY="..."
> export ANTHROPIC_API_KEY="..."

You can customize your OpenAI model and base url by adding the following to your environment:

> export OPENAI_MODEL="..." # default to "gpt-4o"
> export OPENAI_BASE_URL="..." # default to None

You can also use a local model with Ollama. Just add the model name that's being served to your environment:

> export OLLAMA_MODEL="..."

Usage

wut must be used inside a tmux or screen session to capture the last command's output. To use it, just type wut after running a command:

> git create-pr
git: 'create-pr' is not a git command.
> wut

You'll quickly get a brief explanation of the issue:

This error occurs because Git doesn't have a built-in `create-pr` command.
To create a pull request, you typically need to:

1. Push your branch to the remote repository
2. Use the GitHub web interface

If you have a specific question about your last command, you can include a query:

> brew install pip
...
> wut "how do i add this to my PATH variable?"

Roadmap

1. If possible, drop the requirement of being inside a tmux or screen session.
2. Add a --fix option to automatically execute a command suggested by wut.
3. Add wut to Homebrew.
4. Make some unit tests.