Configuration
Strigi is designed to work out of the box without requiring configuration. However, for curated Strigi's behavior, there are a few configuration options available.
Changing the AI Model
Strigi uses Google's Gemini models for generating commands and explanations. You can change the model in several ways:
Method 1: Using the Personalize Command (Recommended)
The easiest way to change your default model is through the personalize command:
$ s personalizeWhen prompted with "Which model would you like to use?", select your preferred model from the available options. This will set your default model for all future interactions.
Method 2: Using the Model Flag (Temporary)
You can specify a different model for a single command using the -m or --model flag:
# Use a specific model for one command
$ s -m gemini-2.5-flash-lite "list files in current directory"
# Or with any subcommand
$ s generate -m gemini-2.5-flash-lite "update packages"
$ s explain -m gemini-2.5-flash-lite "git status"
$ s revise -m gemini-2.5-flash-lite "ls -la"Available Models
Strigi automatically fetches the latest available models from Google's API. Common models include:
gemini-2.5-flash(default) - Fast and efficient for most tasksgemini-2.5-flash-lite- Even faster, good for simple commandsgemini-2.5-pro- More capable for complex reasoning
To see all available models, run the personalize command and browse the selection menu.
Model Selection Tips
- For speed: Choose
gemini-2.5-flash-litefor quick command generation - For accuracy: Use
gemini-2.5-flash(default) for balanced performance - For complexity: Select
gemini-2.5-profor advanced tasks
Your model preference is saved in your configuration and will be used for all future commands unless you override it with the -m flag.
Providing General Contextual Instructions
Strigi strives to be adaptable, and you can leverage the s personalize command to provide specific instructions or preferences that influence its responses.
Use the following command to begin the process:
bash$ s personalizeStrigi will prompt you with three questions:
Which model would you like to use?- Choose your preferred Gemini model from the available options.What would you like the model to know about you to provide better responses?- This allows you to share details about your environment or preferences (e.g., "I primarily use a Linux environment"). This information can help Strigi tailor its responses and command suggestions to your specific context.How would you like the model to respond?- Here, you can specify your preferred response style (e.g., brief, detailed, specific tone).
Setting Your Preferred Explanation Language
To initiate the customization process, use the command:
bash$ s personalizeIn the first prompt (about yourself), you can leave the default behavior and press
Enter.In the second prompt (regarding model responses), add a statement indicating your preferred locale. For example:
Respond using pt-br.
Tailoring Command Generation to Your Operating System
To instruct Strigi to prioritize commands specific to your operating system (e.g., Linux distribution), type the following command.
bash$ s personalizeIn the first prompt (about yourself), add a statement indicating your operating system. For example:
I use Arch btw.In the second prompt (regarding model responses), you can leave the default behavior.