3.1 Create a New Prompty

The GitHub Codespaces environment is preconfigured with Prompty tooling to help us create, convert, and run, Prompty assets. Let's start by creating an empty sandbox/ folder to start our ideation from scratch.

---

1. Create Sandbox Folder

1. Open the VS Code terminal in GitHub Codespaces.
2. Run this command to create a new empty sandbox folder:

   mkdir sandbox
   

3. Change to the directory in the terminal.

   cd sandbox
   

---

2. Create New Prompty

As described earlier Prompty is an asset format (.prompty) that contains the prompt template (content) and model configuration metadata (frontmatter) in a single language-agnostic file! The Prompty runtime (built into VS Code) executes the asset by rendering the template (using the sample data) and invoking the configured model with this enhanced prompt - allowing you to test and iterate rapidly to ideate your application. Explore Prompty documentation for more details.

1. Open the VS Code "File Explorer" view and locate the sandbox/ folder

2. Click on it to get a dropdown menu - look for the New Prompty option shown

   

3. Click on it - this creates a basic.prompty file and opens it in VS Code

3. Explore Prompty Asset

1. View the basic.prompty asset in your VS Code Editor.

   (Click to expand) The starter Prompty asset looks like this:

   ---
   name: ExamplePrompt
   description: A prompt that uses context to ground an incoming question
   authors:
       - Seth Juarez
   model:
       api: chat
       configuration:
           type: azure_openai
           azure_endpoint: ${env:AZURE_OPENAI_ENDPOINT}
           azure_deployment: <your-deployment>
           api_version: 2024-07-01-preview
       parameters:
           max_tokens: 3000
   sample:
       firstName: Seth
       context: >
           The Alpine Explorer Tent boasts a detachable divider for privacy, 
           numerous mesh windows and adjustable vents for ventilation, and 
           a waterproof design. It even has a built-in gear loft for storing 
           your outdoor essentials. In short, it's a blend of privacy, comfort, 
           and convenience, making it your second home in the heart of nature!
       question: What can you tell me about your tents?
   ---
   
   system:
   You are an AI assistant who helps people find information. As the assistant, 
   you answer questions briefly, succinctly, and in a personable manner using 
   markdown and even add some personal flair with appropriate emojis.
   
   # Customer
   You are helping {{firstName}} to find answers to their questions.
   Use their name to address them in your responses.
   
   # Context
   Use the following context to provide a more personalized response to {{firstName}}:
   {{context}}
   
   user:
   {{question}}
   

2. You will observe the following:

   1. The asset uses a YAML format with frontmatter (top) and content (bottom)
   2. The frontmatter has app metadata, model parameters, and sample data
   3. The template has a system message, user question, context and instructions

Note how the asset captures all the elements typically found in a model playground, for prompt engineering. All we need is a runtime to execute the prompt by invoking the model.

4. Run The Prompty

The Prompty extension provides a built-in runtime we can use for this purpose. Let's try it!

1. Open the basic.prompty file in the VS Code editor.
2. Click the play icon seen at top-right corner (or press F5).

3. You will be prompted to sign into Azure as shown. Click Allow.

   

4. Complete the auth flow. This allows Prompty to invoke the Azure-deployed model.

5. Prompty executes the request and displays response in the VS Code Terminal.

   NOTE: This step will fail with an error. Don't worry, that's expected.

   Your Visual Studio Code terminal will switch to the "Output" tab and display this message:

   ❌ | Error: 404 The API deployment for this resource does not exist.

We'll debug this error - and start our first iteration of this ideation process, next!

---

CONGRATULATIONS. You created and ran your first Prompty!