5. Exporting and Using Your Model¶
A trained model isn't locked inside Teachable Machine — it can be exported and actually used elsewhere.
Export options¶
| Format | Used for |
|---|---|
| TensorFlow.js | Running the model in a web page, directly in a browser |
| TensorFlow (Keras) | Running the model in a full Python environment |
| TensorFlow Lite | Running the model on phones or small embedded devices |
Teachable Machine also generates ready-made sample code for whichever format you choose, and offers a shareable link so someone else can open your exact trained project without you having to export any files at all.
What "using it" actually looks like¶
Most simply, that generated web-page sample code is a working page you can open in a browser right now — upload or point a camera at something, and it runs your model's prediction live. That's a real, complete use of your model, with zero additional setup required.
Try it
Export your Lesson 2 model as TensorFlow.js and open the sample page it generates. Confirm it gives the same predictions you saw in Teachable Machine's own Preview panel.
Tip
The shareable link is the easiest way to let someone else try your exact model without needing to re-train anything themselves — useful for showing your Unit 4 capstone project to someone else.