Your first native-app exercise should teach a small connection you can explain: an event triggers a command and produces an observable response. MIT App Inventor's HelloPurr lesson supplies that shape with an image button and a sound. Reproducing it demonstrates a narrow learning result, rather than an original product or a business.
This guide follows the official written lesson and adds an acceptance method. It does not report a performed device test. Use a supported setup and choose an observable exercise suited to your access needs.
Check the computer and device route first
Open MIT's App Inventor setup instructions. The current page recommends Companion testing with a supported phone or tablet and Wi-Fi. It also describes Android USB and compatible Chromebook alternatives.
As checked November 4, 2026, the emulator route is marked “Under Revision: Not Recommended.” Do not treat it as an equally ready fallback merely because an older tutorial lists it. Check the current instructions and actual device support before choosing your learning route.
Write down the computer, browser, device, connection method, and available assistance. If a school or workplace network prevents the live connection, investigate the documented setup option instead of changing unrelated program blocks.
Define a small observable result
The HelloPurr lesson uses a cat image on a button and a meow sound. Its useful learning question is: “Can I connect this button's click event to this sound component's play command?”
Use the lesson's permitted materials or suitable original media. Keep those files identifiable, and avoid uploading private recordings or images of people without appropriate permission.
If sound is inaccessible or unsuitable for the learner, select a supported alternative with an observable result they can evaluate. Simply adding text beside an uninspected sound does not prove that the sound behavior worked. Ask for a suitable adapted exercise or qualified assistance.
Add components and assign their actual sources
Create a new project. In Designer, add a Button and an instructional Label. Upload the chosen image and set the button's Image property to that file. Add a Sound component and set its Source to the intended uploaded audio.
Keep the component names visible as you learn. For the standard example, Button1 identifies the control and Sound1 identifies the nonvisual component. A Sound component with no source is present in the project but has no intended media to play.
Compare the visible screen with the small brief before writing behavior. If the image is missing, inspect the upload and selected property. Fixing the media reference is a different task from fixing the event handler.
Connect the event to one command
In Blocks, add the when Button1.Click event block. Place call Sound1.Play inside its do area. Explain the relationship aloud or in writing: tapping the button triggers the event, and the command asks the named component to play its assigned source.
Do not add several unrelated features before that relationship is clear. Extra sensors, web calls, and screens make it harder to diagnose the first missing response.
Keep a simple prediction: the image remains the same, while one tap produces the assigned sound. Copying the correct block arrangement is a starting action; predicting and inspecting the result is the learning check.
Observe the connected device result
Use the current documented connection flow and tap the button on the connected device. Compare the response with your prediction. Confirm the project and device are linked to the intended session rather than an earlier project.
Use this record:
| Check | Expected result | If it fails |
|---|---|---|
| Connection | Current project appears on device | Repair setup first |
| Image | Intended uploaded image is shown | Inspect Image property |
| Event | Tapping the button invokes the handler | Inspect block connection |
| Sound | Intended media plays | Inspect Source and device audio |
| One deliberate change | Only the predicted behavior changes | Compare source and active project |
Record actual observations instead of marking the table passed by default. Device volume, output routing, missing media, or a broken live connection can resemble a programming error.
Make one change and explain it
Change the instruction label or replace the permitted image with another appropriate one. Predict what should stay the same: the click still invokes the same sound command unless you deliberately change that mapping.
Then make a second small change only if you can explain the first. Preserve a working version before experimentation, and record the action that restored it. A learner who can explain and reproduce the result has evidence of that limited capability.
For a hypothetical learner, changing the picture does not change the sound because they modified the button's image rather than the Sound source. That distinction is useful understanding, even when the exercise has no paying customer.
Preserve the project and name its limits
Save or export the source through the current tool's supported route. Keep the media and a short explanation of the event-response mapping. Companion execution, a packaged independent app, and store release are separate stages with different requirements.
Do not describe a copied learning exercise as an original customer solution. An original field-reporting app, for example, needs its own user task, data model, permissions, device conditions, and distribution decision. The cat-button exercise does not supply those systems.
Choose the next learning or product task
If setup blocked the lesson, resolve that prerequisite or choose an accessible learning route. If the event-response idea is clear, define a small original job and the information it needs before expanding the interface.
Use the software asset construction guide to keep capability, customer value, and operating economics separate. For a web record workflow, the no-code permissions guide teaches a different bounded result. When an app eventually has real users, installs and retention require platform-specific definitions rather than a count of features completed.

