How it works 💡
Imagineer generates a starting prompt for every block, using:
Your structure (rows + blocks)
The type of block (text, image, audio, etc.)
The variables you defined (age, theme, tone…)
For every block in your structure:
Imagineer generates a prompt automatically
You preview it
If you want changes, you chat with Nookly in the side panel
When it feels right, you Test the prompt
If needed, click Optimize to make the prompt sharper and more efficient
Think of Imagineer as drafting the prompt, and you as teaching the system your teaching style and requirements.
The Prompt Model
Every prompt Imagineer generates follows this logic structure: WHO → TASK → VARIABLES
Part of the prompt | What Imagineer auto-generates | Why it matters |
WHO | Gives the AI a role or expertise (SLP, teacher, psychologist…) | Ensures the output matches your field’s tone and expectations |
TASK | Tells the AI exactly what to create (format, quantity, style) | Produces consistent structure and output every time |
VARIABLES | Inserts your variable tags into the prompt | Makes the tool flexible and customizable |
Example of an auto-generated prompt for a text block:
You are a kindergarten teacher. Create {Number of Steps} steps for {Routine Theme} using simple language appropriate for {Target Age Group}.
You can customize it further through chat.
🔁 The refine → test → optimize cycle
1️⃣ Refine using Nookly's side chat -> You tell Nookly what you want to adjust. Example:
“The "who" should be a pediatric speech and language pathologist to better produce articulation cards.”
2️⃣ Click Test
This shows what Imagineer would generate with the current prompt + variables. Test answers the question: “Did Imagineer understand me?”
3️⃣ Click Optimize
Optimize edits the prompt itself (not the output) to make it clearer, more efficient, and more reusable. Optimize answers the question: 'Can this prompt be better?'
How to tell if a prompt is “good enough” 🧠
Ask yourself:
Does it set the right expertise? (WHO)
Does it clearly state quantity, format, audience? (TASK)
Does it use variables instead of static content? (VARIABLES)
Will this produce good results even if someone else uses this tool?
Examples of strong WHO prompts:
Use case | Strong WHO prompt |
SLP articulation cards | You are a pediatric SLP who creates engaging articulation materials for early learners. |
Math problems generator | You are a 3rd-grade math teacher skilled at breaking down multi-step problems. |
Behavior routine | You are an ABA therapist trained in visual routines and first/then strategies |
Examples of strong TASK prompts:
Weak task | Strong task |
“Write routine steps.” | Create {Number of Steps} simple steps for {Routine Theme} using 1 short sentence per step, written for {Target Age}. |
“Flashcards.” | Generate {Number of Trials} flashcards using the {Target Sound} in the {Position in Word} position. |
If yes → you’re ready to go! If not → chat with Nookly and refine.
✨ Important Note: You need to click on each individual block to edit and optimize the prompt before clicking on "Preview & share" as you won't be able to go back to your prompts later.
Once you're happy with your prompts, you can now bring your tool to life and your very first sample will start generating!