Reduce latency and extend battery life by running ultra-tiny AI models directly on Nordic SoCs with automated network optimization and NPU acceleration.
Teachable Machine
Train custom machine learning models for images, sounds, and poses in minutes without writing a single line of code. Ideal for educators, students, and makers.
About Teachable Machine
Teachable Machine is a web-based platform developed by Google that democratizes the creation of machine learning models. Its primary purpose is to provide a fast, accessible environment where users can train computers to recognize specific patterns in images, sounds, and body positions. By removing the technical barriers associated with traditional data science, it allows individuals without coding expertise to explore the possibilities of artificial intelligence through a visual, intuitive interface. The tool operates as a Google Experiment, focusing on ease of use and rapid prototyping rather than complex model architecture management. The workflow is divided into three distinct phases: Gather, Train, and Export. During the gathering stage, users categorize examples by uploading files or capturing live data via their webcam or microphone. This raw data is then processed in the training stage, where the model learns to distinguish between the defined classes. Once trained, users can instantly test the model's accuracy within the browser. The final stage allows for the export of these models into various formats, including TensorFlow.js for web applications, or specific files for hardware-based projects like Arduino and Coral. This tool is exceptionally well-suited for educators, students, and creative professionals. In academic settings, it serves as a practical resource for teaching AI ethics and the fundamentals of classification. For artists and makers, it provides a bridge between software logic and physical interaction, enabling projects like gesture-controlled games or sound-responsive installations. It is also a valuable asset for developers who need to quickly validate an ML concept before committing to more intensive development cycles. What sets Teachable Machine apart from other machine learning platforms is its balance of simplicity and portability. It is entirely web-based, meaning there is no software to install, and it emphasizes user privacy by offering on-device processing so that sensitive data never leaves the local machine. Furthermore, the ability to export real TensorFlow.js models means that the outputs are not just toys; they are functional, standardized models that can be integrated into professional-grade sites, apps, and hardware components.
Teachable Machine pros & cons
Pros
- Completely free to use without any coding or expertise required.
- Supports real-time on-device training to ensure user privacy.
- Exports models to standardized TensorFlow.js formats for wide compatibility.
- Provides immediate visual feedback during the training and testing phases.
- Integrates with popular hardware platforms like Arduino and Coral.
Cons
- Limited to three specific categories: images, sounds, and poses.
- Requires a modern web browser and stable hardware for training performance.
- Does not offer deep architectural control over the underlying neural networks.
- Models are primarily intended for classification tasks rather than complex generation.
Teachable Machine use cases
- Educators can lead lessons on AI ethics and bias by showing students how data classification works in real-time.
- Makers can connect trained pose models to Arduino to build physically interactive games and hardware experiments.
- Developers can rapidly prototype image recognition features for web apps without setting up local ML environments.
- Accessibility researchers can build gesture-to-speech tools to help individuals with limited mobility communicate more easily.
- Artists can create interactive installations that respond to specific audience sounds or body movements.
Teachable Machine features
- hardware integration support
- live webcam data capture
- on-device processing option
- tensorflow.js export
- one-click model training
- body pose detection
- audio sample recognition
- image classification training
Teachable Machine pricing
Is Teachable Machine free? Yes, Teachable Machine has a free plan.
Standard
Free
- Image classification
- Sound recognition
- Pose detection
- Web-based training
- TensorFlow.js export
- On-device processing
- Cloud hosting
- No-code interface
Teachable Machine FAQs
Can I use Teachable Machine without writing any code?
Yes, the platform is designed to be completely code-free for the training and testing phases. While you might need some code to integrate the exported model into your own application, the creation process is entirely visual.
Is my webcam and microphone data uploaded to Google servers?
You can use the tool entirely on-device, meaning your images and audio samples never leave your computer. This makes the platform respectful of your privacy and the way you work.
What formats are supported for exporting models?
Models can be exported as TensorFlow.js models that work anywhere JavaScript runs. You can also export to formats compatible with Coral, Arduino, and other hardware-based platforms.
What types of data can I teach the computer to recognize?
You can train models to classify three types of input: images (via webcam or files), sounds (via short recordings), and body poses (using files or live webcam strikes).
Where can I use the models I create?
The exported models are highly flexible and can be used in websites, mobile apps, or hardware projects. They integrate seamlessly with tools like Glitch, P5.js, and Node.js.
Ratings & reviews
No reviews yet. Be the first to share how Teachable Machine worked for you.
Alternatives to Teachable Machine
Build and productionize AI models and agentic applications with confidence using an integrated developer platform for experiment tracking and model management.
Streamline machine learning research with a unified platform for distributed training, experiment tracking, and multimodal fine-tuning across any compute hardware.
NeuroCraft is a platform for designing, training, and deploying neural networks with a drag-and-drop interface and on-demand pricing.
Mage AI helps product developers build ranking models to increase user engagement and retention, by personalizing content, recommending products, and targeting promotions.
Build decentralized large language models and RAG-based chatbots while retaining ownership of your knowledge contributions on an open, community-driven platform.
Build and deploy computer vision models without training data using plain language.
Rapidly develop and deploy custom machine learning models for image and text classification without a PhD, enabling developers to build accurate AI in minutes.
Sponsored tools
Test your knowledge in live multiplayer Jeopardy-style trivia battles with AI-assisted grading that recognizes correct answers even with spelling mistakes.
Connect with engaged niche audiences or monetize your subscriber base through an automated marketplace featuring verified metrics and secure Stripe payments.
Track brand citations and mentions across Google AI Mode, ChatGPT, Claude, and more. Discover top-cited sources and queries with zero signup required.
Monitor brand citations and prompt visibility across Google AI Overviews, ChatGPT, Perplexity, Claude, and Gemini with free 30-day answer engine reports.
Generate cinematic videos with synchronized audio and multi-modal references. Empower creators and marketers to direct consistent characters, action, and sound.
Create 2K AI videos with native stereo audio from text or images. Ideal for creators and marketers seeking rapid, studio-grade visual content production.