Lobe

What Lobe does

Lobe was a San Francisco-based machine learning tools company, founded in 2016 and associated with Microsoft, that built a free desktop application for Mac and PC enabling users without coding expertise to train computer vision and image-classification models through a visual interface. The product's core technical contribution was abstracting model training behind a no-code UI while exporting trained models for deployment to any platform, supported by an ecosystem of open-source SDKs and bootstrap projects for iOS (Swift), web (TypeScript), Python, .NET, and Flask-based REST APIs, as well as a hardware kit co-developed with Adafruit for IoT deployments. The company's GTM was strictly product-led and self-serve: distribution via direct download from lobe.ai and GitHub, with no enterprise sales motion and no paid pricing tier.\n\nLobe raised a funding round from Lowercase Capital in 2018 and was subsequently associated with Microsoft, as evidenced by the [email protected] corporate email and the GitHub organization's placement within the Microsoft ecosystem. Its customer base was a horizontal mix of non-technical "citizen data scientists," mobile and web developers integrating pre-trained models, and hobbyists/makers — with no disclosed enterprise customers and no commercial revenue stream because the desktop application was free.\n\nThe company formally discontinued the Lobe desktop application, announcing via its GitHub profile that the product is no longer under active development. The entity's operating status is effectively dormant: the core product is wound down, open-source repositories remain available but largely in maintenance, and there is no disclosed leadership team, no recent funding, and no commercial roadmap. Lobe is best understood today as a former Microsoft-affiliated no-code ML tool whose capabilities and talent have been absorbed upstream rather than as an operating business.

Name

Lobe

Legal name

Lobe

Website

https://lobe.ai

Company type

Private

Founded year

2016

Operating status

Closed

Headcount range

11–50 employees

Short description

Lobe was a free, no-code desktop application that let non-technical users train and export image-classification machine learning models to iOS, web, Python, and .NET. Founded in 2016 in San Francisco and affiliated with Microsoft, the product has been discontinued.

Ownership category

akta.pro rank

Product category

Machine Learning Development Tools

NAICS

Business Schools and Computer and Management Training (6114)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

Model Development & Training Platforms (AutoML, Notebooks, Feature Stores) (HDAEANAB)

akta.pro secondary industries

On-Device Inference Runtimes & SDKs (mobile/embedded) (HDAAAJAB), Model Deployment, Serving & Inference Platforms (HDAAABAF)

Keywords

• Machine learning tools
• No-code AI
• Visual model training
• ML deployment SDKs
• Computer vision software

Headquarters

HQ city

San Francisco

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2C

Offering type

Software

Cost components

Technology or R&D, Personnel, Infrastructure, Marketing or Sales

Pricing tiers

Go-to-market motion1 record

Distribution channels2 records

Marketing channels3 records

Core offering

Lobe provided a free desktop application for Mac and PC that allowed users to train machine learning models—primarily for image classification/computer vision tasks—through a visual, no-code interface. Trained models could be exported and deployed to any platform via open-source SDKs and starter projects for iOS, web (TypeScript), Python, .NET, and a Flask REST API, along with image dataset preparation tools and an Adafruit hardware kit. The desktop application is no longer under active development, though open-source repositories remain available.

Product overview

Lobe was a machine learning platform consisting of a desktop application (now discontinued) and an ecosystem of SDKs, starter projects, and tools. The core offering included the Lobe desktop application for training ML models, complemented by language-specific SDKs (lobe-python, lobe.NET), platform-specific bootstrap projects (iOS-bootstrap, web-bootstrap), a REST API starter (flask-server), image dataset tools, and an Adafruit hardware kit partnership. The company is associated with Microsoft and is based in San Francisco.

Differentiator

Problem solved

Functional benefit

Products and services

• Lobe Desktop Application Free desktop tool for Mac and PC that lets non-technical users train machine learning models (primarily image classification) through a visual interface and export them to any platform. Discontinued and no longer under active development.
• lobe-python Open-source Python SDK/toolset for programmatic use of Lobe-trained ML models.
• lobe.NET Open-source .NET library that enables integration of Lobe-trained ML models into .NET applications and services.
• iOS-bootstrap Swift starter project for deploying Lobe-trained machine learning models to iOS applications.
• web-bootstrap TypeScript starter project for deploying Lobe-trained machine learning models to web platforms.
• flask-server Flask-based REST API starter project for serving Lobe-trained machine learning models over HTTP.
• image-tools Tools for creating image-based datasets to support ML model training workflows.
• lobe-adafruit-kit Machine learning kit developed with Adafruit that provides everything needed to bring ML ideas to life on hardware/IoT devices.

Segments3 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration4 records

AI capability2 records

Feature3 records

Partnerships

One partnership is on record.

• AdafruitminorStrategic or Co-development PartnerLobe partnered with Adafruit to create a machine learning kit that helps users bring their machine learning ideas to life. The lobe-adafruit-kit repository contains everything needed to get started with ML projects in partnership with Adafruit's hardware ecosystem.

Scale indicators2 records

Recent moves5 records

Expansion highlights2 records

Direct peers

• Roboflow: End-to-end computer vision platform with no-code training, dataset management, and deployment tooling. Overlaps Lobe on the image-classification use case and similar developer/export experience.
• Edge Impulse: No-code platform for training and deploying ML models on edge devices (microcontrollers, mobile, gateways). Directly competes with Lobe's edge-deployment narrative, particularly via its Adafruit-compatible hardware workflows.
• Apple CreateML: Apple's no-code ML training app for macOS that trains vision, NLP, and tabular models and exports to Apple platforms. Closest functional analog to Lobe in the Apple ecosystem, with the same visual-training paradigm.
• Google Teachable Machine: Google's web-based no-code tool for training image, audio, and pose classification models. Directly overlaps Lobe's core value proposition of visual ML training without code, with similar export-to-platform workflows.

Emerging players

• MakeML: No-code ML tool for creating object detection and segmentation datasets and training models. Direct functional overlap with Lobe's image classification workflow, with active development in the niche.
• Obviously.AI: No-code AutoML platform focused on tabular predictive modeling for business users. Shares Lobe's 'ML for non-technical users' thesis but on structured/tabular data rather than images.
• RunwayML: Creative AI platform that lowers the barrier to ML-powered video and image generation/editing. Shares Lobe's design ethos of bringing ML to non-technical creators with a polished UX.
• Peltarion (now part of Ardor): No-code enterprise AI/ML platform for building and deploying deep learning models. Comparable visual paradigm for non-expert model builders, positioned for operational ML rather than desktop consumers.

Others

• Weights & Biases: MLOps platform for experiment tracking, model management, and dataset versioning. Adjacent infrastructure layer that Lobe-style tools rely on for serious model development and deployment at scale.

Broad incumbents

• Microsoft Azure Machine Learning Designer: Microsoft's enterprise drag-and-drop ML pipeline builder that effectively replaced Lobe within Microsoft's own stack. Targets a similar no-code audience but at enterprise scale and with deeper Azure integration.

Market position

Strengths3 records

Weaknesses3 records

Competitive moat3 records

Key risks5 records

Key highlights5 records

Customer concentration

• Xx.com/lobe_ai
• GitHubgithub.com/lobe
• Facebookfacebook.com/lobeai

Revenue estimate

Valuation estimate

Number of profiles

Funding overview

Funding rounds1 record

Investors1 record

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Lobe