Usage
Shell mode vs CLI mode
| Mode | Command | Best for |
|---|---|---|
| Shell mode | tutr | Interactive terminal sessions with wrapper behavior and prompt context. |
| CLI mode | tutr-cli | One-shot command generation and configuration commands without wrapper shell behavior. |
Use tutr when you want ongoing in-terminal assistance. Use tutr-cli when you want direct command generation behavior or to run setup/configuration commands.
Command shape
tutr <command> <what you want to do>
Examples
tutr git "create and switch to a new branch called testing"
tutr sed "replace all instances of 'foo' with 'bar' in myfile.txt"
tutr curl "http://example.com and display all request headers"
CLI flags
| Flag | Description |
|---|---|
-h, --help | Show help |
-V, --version | Show version |
-d, --debug | Enable debug logging |
-e, --explain | Show explanation and source for the generated command |
Configure command
tutr-cli configure
tutr-cli configure --provider openai --model openai/gpt-4o --show-explanation
tutr-cli configure --provider anthropic --model anthropic/claude-sonnet-4-6
tutr-cli configure --provider ollama --ollama-host http://localhost:11434
tutr-cli configure --clear-api-key
Security note: avoid passing secrets with --api-key because command-line arguments can leak through shell history and process listings. Prefer interactive tutr-cli configure prompts or provider API key environment variables.
Data sent to model providers
When tutr calls an LLM API, it may send your natural-language query, command reference context from <command> --help and/or man <command>, basic system information, and in shell wrapper mode up to 2048 characters of recent terminal output after a failed command.
If you use a remote provider, treat this as data sent to that provider and avoid including secrets in commands, queries, or terminal output.