Data & Analytics

Clearbox AI specializes in generating high-quality, privacy-compliant synthetic datasets, enabling organizations to accelerate AI innovation without compromising sensitive information. The platform helps overcome limitations of traditional data, such as scarcity and sensitivity, by providing privacy-safe datasets that comply with regulations like GDPR. Clearbox AI allows users to replace cumbersome anonymization techniques, improve model performance by enriching targeted data segments, and support research initiatives by simulating complex scenarios. It caters to data scientists, innovation managers, compliance officers, and researchers, offering solutions to enhance predictive analytics, customer insights, and data sharing while mitigating privacy risks.

Convr is an AI-powered underwriting workbench designed for the commercial Property & Casualty (P&C) insurance industry. It leverages a commercial P&C insurance ontology to digitize and enrich submissions, providing critical underwriting insights, accurate business classification, and comprehensive risk scoring. The platform integrates advanced AI, real-time data analytics, and automated workflows to transform risk assessment and decision-making, significantly reducing submission-through-quote times by up to 70%. Convr offers modular AI solutions including Intake, Risk 360, Answers, Scores, Insights, and Workflow, supporting carriers, reinsurers, MGAs, and brokers in achieving greater efficiency and accuracy in their underwriting processes.

AiAssistWorks provides AI-powered add-ons for Google Sheets, Docs, and Slides, enabling users to generate content and automate tasks by simply typing what they need. The tool integrates with over 100 AI models, giving users flexibility and access to the latest AI capabilities. It emphasizes ease of use, requiring no complex formulas or technical skills. Users bring their own API keys, ensuring full control over AI usage and costs, and direct access to preferred models. AiAssistWorks offers a free-forever Lite plan for casual users, providing 100 execution credits per month, making it an affordable solution for various data and content generation needs within the Google Workspace ecosystem.

Image2Table is an AI-powered tool designed to extract tabular data from images and convert it into a structured CSV format. This functionality is particularly useful for automating data entry processes and streamlining data analysis from visual sources. The tool leverages machine learning to accurately identify and interpret table structures within images, making it an efficient solution for converting scanned documents, screenshots, or other image-based tables into editable and analyzable data. While the current live status indicates a build error, its core purpose is to provide a free and accessible way to transform visual data into a usable format for various applications.

H1 (formerly Veda) is an AI-powered platform designed to create a healthier future by making healthcare information and evidence-based medicine accessible globally. It helps life sciences, pharma, health plans, and digital health organizations identify key providers and prescribers, accelerate clinical trials, and advance patient care. The platform offers solutions across clinical, medical, commercial, health plans, and digital health domains, providing tools like Site Universe, Patient Universe, HCP Universe, and Prescriber Universe. H1 also offers data solutions such as Intelligence Streams and MDM services to unify and enrich healthcare data, enabling targeted engagements and efficient operations. It is trusted by leading organizations, including 85 of the top 20 pharma companies and 9 of 10 top payers.

Music Tagging is an AI-powered tool designed to automatically predict and tag music genres. This application leverages machine learning to analyze the characteristics of audio files and assign appropriate genre labels. It is particularly useful for tasks related to music information retrieval and analysis, offering a streamlined approach to organizing and understanding musical content. The tool is available as a Hugging Face Space, making it accessible for users interested in exploring AI applications for music categorization. While the live website currently indicates a runtime error, its intended function is to provide efficient and automated music genre tagging.

PARSeq OCR is an AI-powered Optical Character Recognition (OCR) tool available as a Hugging Face Space. It is designed to extract text from various image and document formats, making it suitable for tasks requiring automated text recognition. While the live website indicates a runtime error, suggesting it may not be fully operational at the moment, the tool's purpose is to provide a platform for OCR capabilities. It is particularly useful for developers, researchers, and data scientists who need to process and analyze text embedded within visual media for their projects.

TableDetAndRec is an AI tool designed for detecting and recognizing table structures within images. Users can upload an image containing a table, and the application will process it to extract the embedded data. The extracted information is then rendered in an HTML format, making it easy to integrate or analyze. Additionally, the tool provides a visual representation of the detected table boundaries and OCR (Optical Character Recognition) boxes directly on the original image, offering transparency and verification of the extraction process. This functionality is particularly useful for automating data processing tasks from scanned documents or images.

GeoX Innovations provides high-quality, up-to-date property data on demand by leveraging advanced AI technology. The tool automatically identifies and extracts more than 20 distinct property attributes from aerial imagery, including building footprints, swimming pools, parking lots, skylights, solar panels, roof type, roof condition, and roof material. This detailed data extraction is crucial for various applications within the real estate and insurance sectors. By automating the process of identifying and cataloging these features, GeoX Innovations helps businesses gain valuable insights into properties without manual inspection, streamlining operations and enhancing data-driven decision-making.

Intics provides Agentic Document Intelligence (ADI) to revolutionize document processing by handling 100% of documents, including complex unstructured or handwritten ones. Unlike traditional methods, Intics offers a no-touch ADI system with full autopilot feedback loops, ensuring high accuracy and efficiency. It leverages pre-trained large vision models (Krypton and Radon) for data extraction without the need for additional training. The platform is designed to work across various industries and data types, offering a scalable solution for processing millions of documents. Intics aims to eliminate manual intervention, reduce costs, and provide real-time control over the data extraction process, transforming dormant document assets into actionable intelligence for autonomous enterprises.

Fero Labs provides a Profitable Sustainability Platform designed for process engineers in complex manufacturing industries. It leverages AI-powered diagnostics and process optimization to help engineers identify and resolve production issues significantly faster, mitigate new problems before they impact output, and enhance overall process efficiencies. The platform includes Fero Diagnostics for root cause analysis, Fero Simulator for identifying precise setpoints, Fero Production for 24/7 optimization, and Fero Foundation for data preparation. It helps teams move from investigation to action quickly, reducing trial-and-error changes and maintaining consistent performance. Fero Labs is built for industries like Steel, Chemicals, Oil & Gas, Cement, and CPG, enabling them to build virtual replicas of processes and optimize performance while reducing costs and emissions.

TensorFlow Data Validation (TFDV) is a powerful open-source library designed for exploring and validating machine learning data. It offers highly scalable capabilities for calculating summary statistics of training and test data, integrating seamlessly with a viewer for data distributions and statistics. TFDV automates data-schema generation to define expectations about data, including required values, ranges, and vocabularies, and provides a schema viewer for inspection. A key feature is its anomaly detection system, which identifies issues like missing features, out-of-range values, or incorrect feature types, complemented by an anomalies viewer to help users correct these issues. TFDV is built to work effectively with TensorFlow and TensorFlow Extended (TFX), making it an essential tool for maintaining data quality in ML pipelines.

PyMDE is a Python library designed for computing vector embeddings for finite sets of items, such as images, biological cells, or network nodes. Built with PyTorch, it offers a simple yet general framework called Minimum-Distortion Embedding (MDE), allowing users to easily recreate well-known embeddings or develop new ones tailored to specific applications. PyMDE is competitive in runtime with more specialized embedding methods, with even faster performance on a GPU. It features fast preprocessing routines implemented in Rust, including approximate and exact k-nearest neighbor algorithms and breadth-first search for all-pairs shortest paths. PyMDE can be used to visualize datasets, generate feature vectors for supervised learning, compress high-dimensional data, and draw graphs efficiently.

Archipelago offers an AI agent designed to streamline broker workflows by providing accurate and validated property and casualty data. It addresses the complexities of traditional spreadsheet property schedules, offering solutions for data ingestion, remediation, and recommendations. The platform features an AI agent that runs in the background to resolve issues proactively, and a Hub with power tools to remediate issues, explain impact, and track progress. Archipelago also provides an enterprise-grade platform for value collection, collaboration, and marketing, ensuring scalability, security, and white-glove support for servicing teams, brokers, producers, and analytics teams. It is trusted by leading risk professionals, managing over 1.6 million properties and 2,500 accounts.

torch-template-for-deep-learning is an open-source project providing PyTorch implementations of a wide array of classical backbone Convolutional Neural Networks (CNNs), alongside essential tools for deep learning development. It includes various data enhancement techniques like Cutout and Mixup, a collection of torch loss functions such as Focal Loss and Dice Loss, and numerous attention mechanisms including SE Attention and Self Attention. The template also features deployment modes for PyTorch models, conversion utilities from TensorFlow to PyTorch, and Class Activation Mapping (CAM) methods. This comprehensive resource aims to simplify and accelerate the development of deep learning applications by offering readily available and well-structured components.

Epinote offers a comprehensive suite of services designed to plug people into workflows, specializing in data annotation, data collection, and project support. It enables businesses to save resources by delegating labor-intensive projects and simple tasks to its network of freelancers. The platform provides efficient data annotation for AI & ML, data collection, and various project support functions. Epinote helps companies accelerate go-to-market strategies, streamline back-office operations, and prepare data for AI initiatives. It offers tailored projects with bespoke workflows and supports various departments including sales, marketing, customer support, data, operations, HR, finance, and project management, aiming to double efficiency by replacing manual processes with a mix of technology and on-demand workforce.

ProntoHQ is a comprehensive B2B prospecting platform designed to significantly boost lead generation and sales outreach efficiency. It enables users to build high-converting outreach lists in seconds by finding companies, leads, emails, and phones through over 100 data providers. Key features include finding leads based on persona, waterfall enrichment for higher data accuracy, and real-time tracking of job changes and new hires to identify optimal outreach moments. The platform also offers AI-powered lead qualification and cleaning, integration with popular CRM and outreach tools, and robust data verification processes to ensure high accuracy of contact information. ProntoHQ aims to reduce the time spent on manual list building and improve conversion rates for sales and growth teams.

Similarix is an AI-powered semantic search engine designed to enhance digital asset management within S3 buckets. It goes beyond traditional keyword matching by understanding the context and meaning behind queries, allowing users to search by text or visually similar images. The platform integrates seamlessly with existing S3 storage, adding a thin AI layer for improved search, organization, and optimization. Key features include intelligent AI for making files searchable, semantic search for relevant results, deduplication to manage duplicate assets, and multilingual support for 133 languages. Similarix operates with read-only access to ensure file integrity and offers an API for integration into existing systems. It uses independent AI models, ensuring reliability and data security without reliance on third-party services.

DeepSeek-OCR is an open-source tool developed by DeepSeek-AI, designed for advanced Optical Character Recognition (OCR) with a focus on contexts optical compression. It enables users to explore the boundaries of visual-text compression, offering various resolution modes including native (Tiny, Small, Base, Large) and dynamic (Gundam). The tool is officially supported in upstream vLLM, providing efficient inference capabilities for both image and PDF processing. It also supports inference via Transformers, allowing for flexible integration into existing workflows. DeepSeek-OCR can handle diverse prompts, from converting documents to markdown and free OCR to parsing figures and general image descriptions, making it a versatile solution for developers and data scientists working with visual data extraction.

FAISS (Facebook AI Similarity Search) is a powerful open-source library designed for efficient similarity search and clustering of dense vectors. It offers a wide range of algorithms capable of searching through vector sets of virtually any size, even those that exceed available RAM. Written in C++ with comprehensive Python/numpy wrappers, FAISS also boasts GPU implementations for accelerated performance, making it suitable for large-scale applications. It supports various distance metrics like L2 (Euclidean) and dot product, and can handle cosine similarity. The library provides a flexible index type that allows for trade-offs between search time, search quality, memory usage, and training time, making it adaptable to diverse needs in similarity search and data analysis.

Parsera is a lightweight Python library designed for efficient web scraping using Large Language Models (LLMs). It provides a straightforward interface, allowing developers to easily extract structured data from websites. Users can define the elements they wish to scrape, such as titles, points, or comments, and Parsera will return the data in a JSON format. The library supports both synchronous and asynchronous operations, and can be run via pip installation, Jupyter Notebook, CLI, or Docker. It also offers flexibility to integrate custom LLM models and playwright scripts, making it a versatile tool for data extraction tasks.

similarities is a comprehensive, open-source toolkit designed for advanced similarity calculation and semantic search. Built with Python 3, it offers out-of-the-box functionality for various tasks, including text-to-text, text-to-image, and image-to-image searches, capable of handling billion-level datasets. The toolkit features semantic matching models based on text2vec for text similarity and search, supporting multiple SentenceBERT-like pre-trained models across various languages. It also includes literal matching models like Word2Vec and BM25. For image and cross-modal similarity, similarities leverages CLIP models, enabling image-to-image, text-to-image, and vector-to-image searches with support for Chinese-CLIP models and GPU acceleration. It provides command-line tools for vector extraction, index building, batch retrieval, and service deployment, making it a versatile solution for developers and data scientists.

CNTXT AI provides comprehensive AI solutions and data services tailored for enterprises and governments, specializing in transforming raw data into AI-ready assets. Their offerings include data services for organizing, labeling, and preparing data, ensuring full in-region compliance and validation by Arabic-native experts. They also design and deliver custom AI systems, from rapid pilots to enterprise-wide deployments, optimized for measurable ROI. The AI Product Lab develops domain-specific models and AI-first applications, including Munsit, an accurate Arabic Speech-to-Text model, and TestAI, an AI validation platform. CNTXT AI emphasizes Arabic-first AI excellence, end-to-end AI readiness, and sovereign-hosted solutions, ensuring data security and compliance with local frameworks.

Infactory is a specialized tool designed to convert various forms of content, including articles, raw data, and extensive content archives, into formats that are optimized for AI systems. This platform empowers users to efficiently query, accurately cite, and effectively license their digital assets. Its primary goal is to assist content creators and data providers in monetizing their intellectual property by making it readily accessible and usable by artificial intelligence. By focusing on transforming content into AI-compatible formats, Infactory streamlines the process of integrating existing knowledge bases with advanced AI applications, opening new avenues for content utilization and revenue generation.