How to Build an AI Microapp

You can combine phases, fields and prompts to make more complex apps...

Video Demo: Moderate Complexity App

Fields

Fields collect information from the user, and they use standard fields types like text input, checkboxes and radio buttons, dropdowns, etc. Each field has a label, a name, and an optional description.

• Label is the text that appears above the field, like the question.
• Name Name is a short, sensible name for the field. You use this name when you add the field to your prompt.
• Description is extra descriptive text for your user.

Prompts

Prompts are how you tell the AI what to do. There are three kinds of prompts that you can use:.

• Prompts are standard text prompts, like you would use in a chatbot. Except that with MicroApps you can include placeholders for fields.
• AI Instructions are instructions for that specific step that you want the AI to follow. Use instructions when you want to give the AI specific instructions for a step.
• Fixed Responses are hard-coded responses that you want to output. The response will be exactly as you write it. You can use field placeholders here too.

Finally, you can include logic like field validation, conditional logic, and scoring to really control how you want the AI to behave based on your inputs.

Video Demo: Advanced App

Advanced Logic

You can build really precise apps by using a combination of field validation, conditional logic, and scored phases:

• Field Validation allows you to control whether or not a field is required, and what kind of input restrictions it has, like a minimum character count.
• Conditional Logic allows you include or exclude a field based on the value of another field. Conditional logic works with both input fields and prompts.
• Scored Phases use a rubric to score the user's response. Responses that don't meet the minimum score threshold will not be allowed to advance.