Feature Request: Teachable Machine Integration with micro:bit in MakeCode (Proposal from an Educator)

playa2021git · June 2, 2025, 2:54pm

Hello everyone,

I am a technology education teacher from Japan, and I am reaching out to propose a new feature for MakeCode: the integration of Teachable Machine models with micro:bit.

In my classes, we are actively exploring “reinvention” projects that combine Micro:bit, generative AI, and machine learning cameras. I strongly believe that MakeCode’s intuitive block-based programming environment is crucial for introducing AI concepts to students at an early age.

Having the ability to easily load and utilize Teachable Machine models directly within MakeCode would be incredibly valuable for fostering students’ creativity and enabling them to build truly innovative AI-powered creations without needing complex external tools. This would greatly benefit AI education in schools.

I have opened a GitHub Issue for this feature request to formally propose it to the MakeCode development team. If you find this idea valuable or have any thoughts to share, I would greatly appreciate it if you could add your support or comments on the GitHub Issue directly. Your collective voices can help demonstrate the demand for this feature.

Link to the GitHub Issue:

github.com/microsoft/pxt

Feature Request: Teachable Machine Integration for micro:bit

opened 12:59AM - 02 Jun 25 UTC

playa2021git

### What problem does this solve? Currently, there is no direct and straightfor…ward way to integrate Teachable Machine models with micro:bit within the MakeCode environment. This limitation prevents students from easily exploring and creating advanced AI/ML projects using the accessible MakeCode platform. As a technology education teacher, I believe that a seamless integration would significantly enhance AI education for younger learners, fostering their creativity and problem-solving skills by allowing them to quickly prototype AI-powered devices. ### Describe the solution you'd like I propose the development of a MakeCode extension that enables direct integration with Teachable Machine models. This extension should allow users to: * Load a trained Teachable Machine model (e.g., via a URL). * Receive input from a camera connected to the micro:bit (perhaps through existing or future AI camera extensions for micro:bit). * Perform inference using the loaded model. * Access the prediction results (e.g., recognized class or confidence scores) within MakeCode blocks. * Trigger actions on the micro:bit based on these predictions. Ideally, the extension would provide user-friendly blocks such as: * `load Teachable Machine model from [URL]` * `get Teachable Machine prediction result` * `if Teachable Machine recognizes [class name] then [do something]` Offline capability would be highly beneficial for classroom environments, but even an online-dependent solution would be a significant step forward. Support for other sensor inputs like microphones for audio recognition would also be a great addition in the future. ### Describe alternatives you've considered Currently, advanced machine learning projects involving micro:bit typically require external tools like Python on a PC, which adds complexity and makes it less accessible for beginners or in a block-based coding environment like MakeCode. This approach deviates significantly from MakeCode's philosophy of simple and intuitive programming. ### Additional context As a technology education teacher in Japan, I am passionate about integrating cutting-edge technology, including AI, into my curriculum. While my personal coding skills are at a beginner level, I effectively use generative AI and understand the nuances of prompt engineering to develop educational projects. In my classes, we are exploring "reinvention" projects that combine micro:bit, generative AI, and machine learning cameras. I strongly believe that MakeCode's intuitive block-based programming environment is crucial for introducing AI concepts to students at an early age. An extension that facilitates Teachable Machine integration would be invaluable for my teaching and for the broader educational community, empowering students to build truly innovative AI-powered creations. Thank you for considering this request.

Thank you for considering this proposal and for your support!

bsiever · June 3, 2025, 2:41am

Hi @playa2021git & welcome to the forum,

There are several AI integrations, including some that utilize Teachable Machine:

CreateAI
Make AI: Robots site, which is a Teachable Machine integration. Last year they added some chatbot support too. (micro:bit needs to be connected to a computer with a USB cable)
- At some point Techmonstah was interested in a wireless (bluetooth) version of Make AI’s site. Here’s a version of the site that uses bluetooth instead of a USB cable: https://bsiever.github.io/makeairobots/

The awesome Awesome micro:bit list also lists several AI/Machine Learning resources.

playa2021git · June 3, 2025, 2:53am

Thank you for your interest in this feature request.
Please note that the GitHub Issue has been moved to a new address:

github.com/microsoft/pxt-microbit

Feature Request: Teachable Machine Integration with micro:bit in MakeCode (Proposal from an Educator)

opened 09:36PM - 02 Jun 25 UTC

playa2021git

Hello everyone, I am a technology education teacher from Japan, and I am reachi…ng out to propose a new feature for MakeCode: the integration of Teachable Machine models with micro:bit. Last year, I led my students through "reinvention" projects using Streatch3—which incorporated TM2Scratch and Microbit More extensions—as well as coding support from ChatGPT. However, because Scratch only allows block-based programming, and its workflow is not well-matched with AI-powered coding assistants like ChatGPT, most students were unable to fully realize their inventions. The limitations of the environment made it difficult to bridge the gap between creative ideas and technical implementation. In my classes, we are actively exploring projects that combine Micro:bit, generative AI, and machine learning cameras. I strongly believe that MakeCode’s intuitive block-based programming environment is crucial for introducing AI concepts to students at an early age. Unlike Scratch, MakeCode allows seamless switching between blocks and JavaScript, which means students can get the most out of AI-powered coding assistance. Having the ability to easily load and utilize Teachable Machine models directly within MakeCode would be incredibly valuable for fostering students' creativity and enabling them to build truly innovative AI-powered creations without needing complex external tools. This would greatly benefit AI education in schools. I have opened a GitHub Issue for this feature request to formally propose it to the MakeCode development team. If you are a professional programmer or engineer reading this and find the idea interesting, I would be especially grateful for your interest or support. If you have any thoughts to share, I would greatly appreciate it if you could add your support or comments on the GitHub Issue directly. Your collective voices can help demonstrate the demand for this feature. Thank you for considering this proposal and for your support!

I appreciate your continued support and feedback!

playa2021git · June 4, 2025, 5:01pm

Thank you so much for your kind and detailed response!
The links and explanations you shared are extremely helpful.
I tried https://makeairobots.com/ and was able to confirm that it actually links with Teachable Machine—thank you for introducing this!
I especially appreciate the information about the Bluetooth version of the Make AI: Robots site and the resources in the Awesome micro:bit list.
This will greatly help me in exploring AI and machine learning integrations with micro:bit in my classroom.
Thank you again for taking the time to support me and the community!

Topic	Replies	Views
https://makecode.microbit.org/_fRRUpKdPuXKv micro:bit	398	August 1, 2020
Video: Radio with CreateAI Micro:bit micro:bit	33	October 24, 2024
dragonflies_aalyah micro:bit	213	May 17, 2021
testCase micro:bit	195	August 21, 2020
james micro:bit	190	March 12, 2021

Feature Request: Teachable Machine Integration with micro:bit in MakeCode (Proposal from an Educator)

Related topics