lunit

Browser AI assistant for reading, analyzing, and generating text within web pages

A desktop work log tool that captures thoughts and decisions while maintaining a knowledge graph.

Road maintenance robot that automatically detects and seals asphalt cracks

PDF information processor that extracts content to generate summaries, tables of contents, and answer targeted questions.

An academic writing assistant platform supporting reading, writing, citation, and multilingual content generation.

Audio and video transcription tool that converts content into summaries, mind maps, and key Q&A.

The Segment Anything Model (SAM) is a cutting-edge image segmentation model developed by Meta AI's Fundamental Artificial Intelligence Research (FAIR) laboratory. Trained on the largest segmentation dataset to date, comprising over 11 million images and 1 billion masks, the model can generate precise segmentation masks based on various prompts such as points, boxes, or text. Its architecture consists of an image encoder, a prompt encoder, and a lightweight mask decoder, enabling real-time mask generation and strong zero-shot performance across diverse segmentation tasks without requiring additional training. By simplifying the annotation process, SAM supports a wide range of applications—from medical imaging to environmental monitoring—and advances the democratization of image segmentation.

SkinSort is a trusted skincare companion designed to help users make informed skincare decisions. It features powerful ingredient analysis tools that interpret product formulations based on scientific research, enabling users to identify ingredients that benefit or harm their skin. The platform supports scanning product labels, side-by-side comparison of skincare products, finding affordable alternatives, and long-term tracking of skincare routines. SkinSort also fosters a supportive community where users can share routines, give reviews, and offer advice. The platform is independent of brand influence, ensuring neutral and transparent information for all skincare enthusiasts.

CVAT (Computer Vision Annotation Tool) is a comprehensive data annotation platform designed to streamline the process of creating labeled datasets for machine learning and computer vision applications. The platform provides teams with an intuitive interface and powerful annotation capabilities to handle projects of any scale, from small initiatives to enterprise-level datasets. CVAT supports multiple annotation types, including bounding boxes, polygons, keypoints, skeletons, cuboids, and trajectories, enabling users to prepare data for a wide range of machine learning tasks. By integrating semi-automated annotation tools and supporting major cloud storage providers, CVAT accelerates annotation workflows while maintaining data privacy and security.

Picarta AI is a cutting-edge image geolocation platform that uses artificial intelligence to determine the geographic location of any photo worldwide. By analyzing visual features and metadata such as EXIF, Picarta delivers highly accurate GPS coordinates and location predictions. The platform supports multiple access methods, including a web interface, API integration, and enterprise-grade offline models. Picarta's mission is to ensure authenticity, combat misinformation, and provide users with reliable image location verification capabilities.

Pearl is a global leader in dental AI platforms, leveraging advanced computer vision and machine learning technologies to support dental practices worldwide. The platform offers a suite of AI-powered tools designed to enhance diagnostic accuracy, optimize clinical performance, and drive practice growth. Pearl's flagship product, Second Opinion®, enables real-time detection of dental pathologies at the point of care, improving patient trust and treatment planning. Complementary solutions like Practice Intelligence® boost operational efficiency and financial performance by analyzing clinical data. Pearl seamlessly integrates with practice management systems, delivering actionable insights and AI-driven automation to dentists.

Landing AI, founded by AI pioneer Andrew Ng, focuses on visual AI solutions and drives industry transformation by unlocking the value of visual data. Its flagship product, LandingLens, is a cloud-based computer vision platform that enables users to rapidly build, train, and deploy deep learning models for defect detection, classification, and segmentation—regardless of their AI expertise. The platform emphasizes data quality and consistency through advanced annotation tools and a data-centric AI approach, helping enterprises quickly move AI projects from proof of concept to large-scale production. Landing AI supports flexible deployment options, including cloud, edge devices, and integration with third-party platforms such as Snowflake, making it suitable for operations of all scales—from a single production line to global deployments.