AI Machine Learning Tool

The Rise and Quiet Demise of Lobe.ai: Why Microsoft's No-Code AI Tool Vanished
Ever wondered what happened to Lobe.ai, the promising no-code AI tool that Microsoft acquired in 2018? Despite generating excitement for its user-friendly approach to machine learning model training, the platform quietly faded into obscurity, officially discontinued by early 2023. But why?
Did Microsoft intentionally shelve it? Did the technology become redundant? Or was it simply a strategic decision to focus on web-based AI tools? Let's break it all down.
What Was Lobe.ai?
Lobe.ai was a machine learning tool designed to make AI model creation accessible to non-programmers. Its standout feature was a simple drag-and-drop interface that allowed users to train AI models—especially for image classification—without writing a single line of code.
Key Facts About Lobe.ai
- Founded: 2015
- Founders: Mike Matas, Adam Menges, and Markus Beissinger
- Acquired By: Microsoft in September 2018
- Goal: Democratize AI development by offering an intuitive desktop app
- Killer Feature: Local desktop processing for machine learning models, allowing users to train models without cloud dependency
At the time of Microsoft's acquisition, Lobe was positioned as a game-changer for newcomers to AI. But within a few years, it disappeared.
Why Did Lobe.ai Fail?
Short Answer:
Lobe.ai was silently phased out because Microsoft pivoted towards integrated, web-based AI tools—making the standalone desktop application redundant.
Long Answer:
A combination of strategic shifts, technological redundancy, and competition sealed Lobe's fate. Here’s how:
1. Microsoft’s Strategic Shift to Web-Based AI Tools
One of the biggest factors behind Lobe’s shutdown was Microsoft's pivot away from standalone AI apps in favor of integrated, cloud-based solutions like AI Builder.
- In early 2021, Lobe's GitHub page acknowledged that the desktop app was no longer in development, hinting at its eventual end.
- By January 2023, Microsoft officially deprecated Lobe's image classification model, stating it was "not aligned with other models in AI Builder", which was a fully web-based AI training tool.
Essentially, Microsoft decided that web-based AI tools were more scalable and easier to maintain than standalone desktop applications.
2. Technological Obsolescence: Shift to Cloud AI
When Lobe gained traction, local model training was a selling point. Users loved the ability to train AI models on their computers without relying on cloud-based servers.
However, by the early 2020s, cloud AI services were the industry norm. Startups and enterprises were moving towards collaborative, real-time AI training environments hosted in the cloud.
Lobe suddenly looked outdated compared to competitors like:
- Google's AutoML (cloud-based AI model training)
- Microsoft’s own Custom Vision and AI Builder
- Amazon SageMaker (cloud-first machine learning)
Rather than invest in keeping Lobe relevant, Microsoft likely absorbed its best features into its other AI products.
3. Market Competition Made Lobe Expendable
The no-code AI space became crowded with players like:
- Edge Impulse – focused on low-code AI for embedded systems
- Google’s Teachable Machine – similar drag-and-drop AI with broader use cases
- Apple's Create ML – offering machine learning model training for macOS users
Unlike Microsoft’s enterprise-scale AI tools, Lobe lacked a clear path to profitability. It appealed more to hobbyists and small businesses rather than large companies willing to pay for AI solutions—making it an unlikely priority for Microsoft.
4. Microsoft's Resource Allocation Post-Acquisition
Microsoft acquires companies for two main reasons:
- To integrate their technology into Microsoft products
- To acquire talent for internal research and development
Post-acquisition, the Lobe team likely merged into Microsoft's AI projects, helping improve platforms like Azure AI, AI Builder, and Custom Vision.
Meanwhile, maintaining a standalone Lobe-branded product cost Microsoft resources that it could allocate elsewhere. Rather than maintain an isolated desktop app, Microsoft prioritized enterprise-ready cloud offerings.
5. Lack of Monetization Strategy
At its core, Lobe was a free tool with no clear business model.
- Microsoft never introduced a premium version or subscription tiers.
- Without a revenue stream, continuing development likely became financially unviable.
- The return on investment was likely too low to justify maintaining the standalone app.
A tool like Lobe primarily served as an “onboarding” experience for AI development, but Microsoft already had a suite of AI platforms. Lobe simply didn’t fit into the larger Microsoft ecosystem in a way that generated revenue.
Could Lobe.ai Have Survived?
Possibly, but it would have required:
- A pivot to cloud-based AI training rather than a desktop-only model
- Integration with Azure AI services rather than existing as a separate entity
- A strong monetization model (e.g., premium features for enterprise users)
Since Microsoft already had Custom Vision, AI Builder, and Azure ML, continuing Lobe as a separate project made little financial sense.
Final Thoughts: Lessons from Lobe’s Shutdown
Lobe.ai was a brilliant concept that fell victim to the rapid evolution of AI development tools.
Key Takeaways:
- Strategic alignment matters: If a product doesn’t fit into a company's bigger picture, it won’t last.
- Cloud-first AI won: Cloud-based AI modeling became the industry standard, leaving Lobe's desktop approach obsolete.
- Monetization is critical: Free tools need a long-term way to sustain themselves—or risk discontinuation.
- Acquisitions don’t always mean success: Getting bought by a tech giant doesn’t guarantee survival—sometimes it signals an eventual phase-out.
While Lobe is gone, its core mission—making AI model creation accessible to everyone—lives on in other Microsoft AI products.
FAQ Section
Who founded Lobe.ai?
Lobe.ai was founded in 2015 by Mike Matas, Adam Menges, and Markus Beissinger.
When did Lobe.ai shut down?
Lobe’s desktop app was no longer under development by early 2021 and its image classification model was officially deprecated in January 2023.
Why did Microsoft shut down Lobe.ai?
Microsoft likely discontinued Lobe because it was redundant next to web-based AI tools like AI Builder and Custom Vision. Maintaining a desktop app no longer aligned with Microsoft’s AI strategy.
Did Microsoft integrate Lobe's technology into other products?
While Microsoft didn’t publicly announce it, Lobe’s AI tech likely influenced updates to AI Builder, Azure AI, and Power Platform tools. Many of its no-code AI principles live on in Microsoft’s broader AI ecosystem.
What were Lobe.ai's main competitors?
Similar no-code AI tools include:
- Google’s Teachable Machine
- Edge Impulse
- Apple's Create ML
- Microsoft’s own Custom Vision & AI Builder
Can I still use Lobe.ai?
No, Lobe is no longer available for download or use. Microsoft deprecated its core functionality in January 2023.
Though Lobe.ai is no more, its legacy in no-code AI democratization continues to influence AI education and accessible machine learning tools today.
Would Lobe have survived if it had pivoted to cloud-based AI? Maybe. But in the fast-moving tech world, great ideas don’t always guarantee long-term sustainability.
What is lobe.ai?
Lobe.ai is a user-friendly, cost-free tool designed for individuals looking to develop and tailor artificial intelligence models without the necessity of coding. This tool enables users to train their models using images, data, or gestures and subsequently export them to a variety of platforms. It's noteworthy that Lobe.ai is a product of Microsoft, and its primary objective is to democratize machine learning, making it accessible and enjoyable for a wide audience.
What are the benefits of lobe.ai?
Lobe.ai offers several noteworthy benefits:
Cost-Free and Private: Users can employ Lobe.ai on their local computer, ensuring privacy by not requiring them to upload their data to the cloud.
User-Friendly Interface: The tool boasts an intuitive visual interface, allowing users to easily label their images, train their models, and observe real-time results.
Powerful Machine Learning: Lobe.ai includes project templates that intelligently select the appropriate machine learning architecture for specific tasks, such as image classification, object detection, or data classification.
Versatile Exporting: Users have the option to export their models to various industry-standard formats and platforms, including TensorFlow, Core ML, and ONNX, making it compatible with a wide range of applications and systems.
How can I get started with Lobe.ai?
To begin your journey with Lobe.ai, you can easily get started by following these straightforward steps:
Download Lobe.ai App: Start by downloading the Lobe.ai application tailored for Mac or PC. You can obtain it either from their official download page or directly from their [GitHub repository].
Launch the Application: After downloading, launch the Lobe.ai app and initiate a new project. Depending on your specific task, you can opt for a template that suits your needs, such as image classification, object detection, or data classification.
Collect and Label Your Data: Begin collecting and labeling your dataset. You can accomplish this by either utilizing your computer's webcam or by dragging and dropping a folder of images from your local storage.
Automated Model Training: Lobe.ai streamlines the model training process, conducting it autonomously on your own computer. As your model trains, you can observe live results and make necessary adjustments.
Utilize Your Model: Once trained, you can use your model with your webcam or by inputting images from your computer. Provide feedback on its predictions to further enhance its performance.
Export Your Model: Finally, Lobe.ai offers the flexibility to export your model to a range of platforms, enabling compatibility with TensorFlow, Core ML, ONNX, and other options, as per your preference.
How much does lobe.ai cost?
Lobe.ai is a no-cost software solution that empowers users to generate and deploy machine learning models without the need for coding skills. It's important to note that there are no concealed charges or fees linked to the application. To get started, you can acquire the app for Mac or PC either from their official download page or directly from their GitHub repository.
What are the limitations of lobe.ai?
While Lobe.ai is a valuable tool for creating and deploying machine learning models without coding, it does come with some limitations that users should consider:
Limited to Image Classification: Currently, Lobe.ai supports only image classification tasks, which means it can assign labels to entire images but not to specific objects or regions within them. It does not support other types of tasks like text or data classification, object detection, or face recognition.
Internet Connection Dependency: Lobe.ai relies on a fast and reliable internet connection for various functions, including downloading and uploading images, model training, and exporting to different platforms. It does not offer offline functionality or support slow networks.
Storage Restrictions: Users should be aware of limited storage space for projects and images. Lobe.ai allows a maximum of 10 projects and up to 200 images per project on your computer. If you exceed these limits, you'll need to delete existing projects or images to make room for new ones.
Lack of Built-in Evaluation: Lobe.ai does not include a built-in evaluation or testing feature to measure the accuracy or performance of your models. Users must manually assess model performance with new images or rely on external tools for evaluation.
No Continuous Learning or Human-in-the-Loop: The tool lacks continuous learning or human-in-the-loop features, meaning that if you wish to make changes or updates to your models based on feedback or new data, you'll need to retrain them from scratch.
These limitations highlight the areas where Lobe.ai may not be suitable for certain tasks or requirements, so users should consider these factors when utilizing the tool.
What is LobeAI and what does it offer?
LobeAI is a user-friendly machine learning tool designed to make the process of training and deploying machine learning models accessible for users without coding expertise. It offers a variety of benefits including a cost-free usage model, privacy through local data processing, and intuitive visual interface. It supports exporting models to industry-standard formats such as TensorFlow, Core ML, and ONNX, enabling compatibility with various applications.
How can developers integrate LobeAI with different platforms?
Developers can integrate LobeAI with different platforms using the starter projects available for iOS, Android, and web. These projects provide a starting point to bootstrap Lobe machine learning models into mobile and web applications. LobeAI also provides tools like lobe-python for working with Python, offering flexibility for developers to integrate Lobe models into their preferred environments.
What are some popular repositories within LobeAI's GitHub organization?
Within LobeAI's GitHub organization, popular repositories include lobe-python for working with Lobe models using Python, iOS-bootstrap for integrating Lobe models into iOS apps, and web-bootstrap for web applications. Additionally, image-tools are available for creating image-based datasets, and the lobe-adafruit-kit enables bringing machine learning ideas to life with the Adafruit kit. These repositories serve as resources for developers looking to leverage LobeAI in various projects.