Turn plain English into shell commands using OpenAI, Claude, or local LLMs, designed to teach, not replace your knowledge.
We’ve all been there: you know what command you want to run, but the syntax, awk, find, or sed slips your mind. oneliner helps you figure it out in your terminal, so you can learn as you go, without leaving the shell or installing heavyweight tools like Warp or Claude CLI.
Requires Go 1.25.1+
go install github.com/dorochadev/oneliner@latestOr build from source:
git clone https://github.com/dorochadev/oneliner.git
cd oneliner
go build -o oneliner .Run setup:
oneliner setupGenerate commands:
oneliner "find all jpg files larger than 10MB"
oneliner --explain "delete node_modules recursively"
oneliner --clipboard "compress all pdfs"
oneliner --breakdown "list all active network connections with details"Commands are shown, not executed by default. Use
--runonly when you’re sure.
For configuration details, see the Configuration section below.
- Supports OpenAI, Claude, and local LLMs
- Context-aware (OS, shell, directory)
- Pretty terminal UI (Lipgloss & Bubble Tea)
- Fast, cached results
- Clipboard copy, explanations, and detailed command breakdowns
oneliner never runs commands unless you explicitly use --run.
A regex-based safety check warns about dangerous commands, but do not rely on it blindly.
Use
--runand--sudoonly when 100% sure what the command does.
| Flag | Short | Description |
|---|---|---|
--run |
-r |
Execute the command immediately |
--sudo |
Prepend sudo (Unix only) |
|
--explain |
-e |
Show a brief explanation of the command |
--clipboard |
-c |
Copy command to clipboard |
--interactive |
-i |
Run in interactive mode |
--breakdown |
-b |
Full educational breakdown of command stages |
--config |
Use a custom configuration file |
The --breakdown (-b) flag is designed for learning and understanding, not just quick answers. It provides a detailed, step-by-step explanation of how a shell command works.
When you use --breakdown, oneliner will:
-
Show the command first (no execution unless
--runis also used). -
Explain every stage in a numbered list:
- Each pipe, flag, or redirection is explained.
- Data transformations between stages are described.
- Shell expansions, substitutions, or environment effects are clarified.
-
Provide multiple points as needed — there’s no limit, so even complex commands are fully unpacked.
-
Serve as an educational tool so you can learn why and how commands work, not just copy them.
oneliner --breakdown "find all jpg files larger than 10MB"Output will include a numbered breakdown explaining each stage of the
findcommand, what each option does, and how the pipeline processes files.
Manage your LLM setup in one place:
- Interactive Setup:
oneliner setup- View Current Config:
oneliner config list- Set Config Manually:
oneliner config set llm_api openai
oneliner config set api_key sk-xxxx
oneliner config set model gpt-4o- Local LLM Example:
oneliner config set llm_api local
oneliner config set local_llm_endpoint "http://localhost:8000/v1/completions"
oneliner config set model llama3- Config File:
~/.config/oneliner/config.json
oneliner cache list
oneliner cache clear
oneliner cache rm <id>| Issue | Solution |
|---|---|
oneliner: command not found |
Add $(go env GOPATH)/bin to PATH |
| Configuration incomplete | Run oneliner setup |
| API errors | Check API key and connectivity |
| Cache issues | Run oneliner cache clear |
Pull requests are welcome! Open issues for bugs or feature requests.
MIT
Built with ❤️ using Cobra, Bubble Tea, and Lipgloss