Ask DSL: Commands & Parameters

Learn to control Ask behavior with commands and parameters for powerful conversational AI.

What is Ask DSL?

Ask DSL (Domain Specific Language) provides special commands and parameters that give you precise control over how an Ask responds. Think of it as keyboard shortcuts for AI conversations.

Basic Commands

@model - Switch AI Models

Change the underlying AI model mid-conversation:

@model claude-sonnet
Now respond to my question...

Available models:

@model claude-sonnet - Balanced, excellent writing
@model gpt-4o - Strong reasoning
@model gemini-flash - Fast responses
@model deepseek - Cost-effective

/image - Generate Images

Create images inline in conversation:

/image A serene mountain landscape at sunset, digital art style

Image options:

/image:flux - Use FLUX model
/image:ideogram - Use Ideogram (great for text)
/image:4 - Generate 4 variations

/audio - Text to Speech

Convert text to audio:

/audio Read this paragraph in a professional tone

/video - Generate Video

Create short video clips:

/video A cat playing with a ball of yarn, 5 seconds

Parameters ($)

Customize generation with dollar-sign parameters:

Temperature

$temp:0.3 Write a formal business email
$temp:1.5 Write a creative story

Length Control

$max:100 Summarize this in under 100 words
$max:500 Write a detailed explanation

Style Modifiers

$tone:professional Write a response
$tone:casual Explain this concept
$style:bullet List the key points

Knowledge Bases (@kb)

Reference your uploaded knowledge:

@kb:brand-guidelines Create a tagline for our new product
@kb:product-specs Compare our features to competitors

Upload knowledge bases in Ask Settings to make them available.

Combining Commands

Chain commands for complex outputs:

@model claude-sonnet $tone:professional
Write a product description for a wireless speaker

/image:ideogram Product photo of modern wireless speaker, minimalist white background

Inline Media

Reference images in conversation:

[image: url-to-image]
Describe what you see in this image

Or upload directly in the chat interface.

Best Practices

Start simple: Learn one command at a time
Combine wisely: Don't overload with parameters
Use knowledge bases: Better context = better results
Iterate: Refine your prompts based on outputs

Quick Reference

Command	Purpose	Example
@model	Switch AI	@model gpt-4o
/image	Generate image	/image sunset
/audio	Text to speech	/audio Read this
/video	Create video	/video walking cat
$temp	Set creativity	$temp:0.8
$max	Limit length	$max:200
@kb	Use knowledge	@kb:docs

Ready to practice? Find an Ask to try