AI Agents & Automation

Unsiloed AI is an API-native document intelligence tool designed to convert multimodal unstructured data into structured, LLM-ready formats with high accuracy. It addresses the challenge of unstructured data hindering AI adoption by providing advanced vision models for parsing, extraction, and hierarchical splitting. The tool can process various document types including PDFs, images, spreadsheets, and scanned documents, handling complex layouts like tables, charts, and handwritten content. It generates clean, LLM-ready Markdown and structured JSON outputs with confidence scores, and supports schema-validated extractions. Unsiloed AI offers both managed and air-gapped deployment options, ensuring flexibility for enterprise needs.

Spleenlab specializes in developing next-generation AI software for autonomous systems across various industries, including aviation, automotive, defense, and robotics. Their solutions are designed to enable safe, scalable, and certifiable automation. For drones, Spleenlab offers capabilities such as GPS-denied navigation, surveillance, collision avoidance, precision landing, and terrain-following. For vehicles, their AI software supports multi-object detection, sensor fusion, freespace detection, occupancy grids, and GPS-denied SLAM. The company's product suite, VISIONAIRY®, includes modules for camera calibration, spatial detection, object detection, navigation, and localization, all compatible with various hardware, chips, and sensors. SPLEENLAB® Suite further provides video stabilization, an operating system, AI runtime, and simulation tools, ensuring efficient performance with minimal computational resources.

wav2letter is an open-source automatic speech recognition (ASR) toolkit developed by Facebook AI Research. It is specifically designed for AI researchers and speech recognition developers, offering a flexible framework for building and experimenting with ASR models. The toolkit has been consolidated into Flashlight in the ASR application, indicating its integration into a broader machine learning library. While the provided website content is a GitHub pricing page, the context from the tool's description suggests its primary function is to provide foundational tools for advanced speech recognition development, rather than being a consumer-facing application. Users can leverage wav2letter for tasks such as training custom speech models and conducting research in the field of automatic speech recognition.

Awesome-RL-for-LRMs is an open-source project offering a comprehensive survey of reinforcement learning (RL) techniques specifically applied to large reasoning models (LRMs). This resource is invaluable for researchers and engineers looking to understand and implement RL in the context of large language models (LLMs) and other reasoning models. It compiles relevant papers, resources, and insights, making it easier to navigate this complex and rapidly evolving field. The project aims to provide a foundational understanding and practical guidance for those involved in AI model training and development, particularly in areas requiring advanced reasoning capabilities.

VirtualHome is an interactive platform and API designed for simulating complex household activities using programs. It enables multi-agent simulations where agents can interact with environments by picking up objects, switching appliances, and more. The platform offers two simulators: the Unity Simulator for generating videos of activities and the Evolving Graph simulator for tracking environmental changes. It supports various ground-truth streams like time-stamped actions, segmentation, and optical flow, making it suitable for training agents in embodied AI tasks. VirtualHome also features procedural generation for unique environments, enhanced physics, time management, and improved lighting, with a Python API for easy integration and control.

CodeFuse-muAgent is an innovative, open-source agent framework driven by a Knowledge Graph (KG) engine, designed to facilitate complex reasoning and online collaboration among multiple AI agents. It leverages LLMs, FunctionCall, and CodeInterpreter technologies, enabling users to orchestrate agents through a canvas-based drag-and-drop interface or simple text commands. The framework supports one-click deployment, including KG-based agent orchestration and Java-based tool registration and management. Key features include EKG Builder for designing virtual teams and semantic nodes, EKG Assets for comprehensive KG Schema design, and EKG Reasoning for flexible, human-guided LLM operations. It also provides visual debugging, end-to-end monitoring, a unified message pooling system for memory management, and an ActionSpace adhering to the Swagger protocol for secure tool execution. This framework has been validated in complex DevOps scenarios within Ant Group.

voice-elements is a Web Component wrapper for the Web Speech API, designed to facilitate both voice recognition (speech to text) and speech synthesis (text to speech) within web applications. Built with Polymer, it offers a simple DOM API for developers to integrate these functionalities. Key features include a `<voice-player>` component for text-to-speech with options for autoplay, accent, and customizable text, along with methods to speak, cancel, pause, and resume audio. The `<voice-recognition>` component provides speech-to-text capabilities, allowing continuous recognition and returning the recognized text. It also includes methods to start, stop, and abort recognition. The tool provides event triggers for various stages of speech synthesis and recognition, such as `onstart`, `onend`, `onerror`, `onpause`, `onresume`, and `onresult`. While offering powerful features, users should note the current limitations in browser support for the Web Speech API.

WebDataset is a Python-based I/O system specifically engineered for both large and small-scale deep learning tasks, providing robust integration with PyTorch. It streamlines data handling by organizing training samples and datasets within tar files, adhering to specific conventions for efficient access. This approach is particularly beneficial for high-performance data loading, reducing I/O bottlenecks during model training. The tool's design focuses on optimizing data pipelines, making it a valuable asset for developers and data scientists working with extensive datasets in machine learning projects. Its emphasis on structured data organization within tar files facilitates scalable and reproducible research.

wllama is a WebAssembly binding for llama.cpp, designed to enable on-browser LLM inference. This tool allows developers to run large language models directly within a web browser using WebAssembly SIMD, eliminating the need for a backend server or a dedicated GPU. It offers comprehensive TypeScript support and provides both high-level APIs for completions and embeddings, as well as low-level APIs for fine-grained control over tokenization, KV cache, and sampling. A key feature is its ability to automatically switch between single-thread and multi-thread builds based on browser support, ensuring optimal performance. Models can be split into smaller files for parallel downloading, improving load times and handling models larger than 2GB. wllama also includes pre-built npm packages and supports custom logging.

voice-assistant-scripts offers a collection of example scripts designed for AI agents built using the Alan AI Platform. These scripts serve as practical demonstrations of how to structure dialogs between users and AI agents, covering various conversational scenarios. Developers can examine these examples to gain insights into conversational AI design and use them as a foundational starting point for crafting their own custom dialog scripts. The repository includes diverse examples such as Bitcoin calculators, calendars, food ordering systems, news assistants, and translators, showcasing the versatility of the Alan AI Platform. It is an invaluable resource for AI creators and developers looking to implement robust and engaging voice assistant functionalities.

AISent specializes in delivering impactful AI solutions for industrial applications, focusing on computer vision and advanced data analysis. Their Industrial Vision offerings leverage complex algorithms and neural networks for image analysis, pattern recognition, and quality control, opening new possibilities in fields like automotive and luxury goods. Industrial Intelligence focuses on unlocking hidden potential in data, optimizing processes, identifying trends, and informing decision-making across various sectors. AISent also provides an Academy with executive, plant operations, and technical courses to educate professionals on AI's strategic and practical applications in industry. Their solutions are tailored for diverse sectors including Food & Beverage, Automation & Machinery, Transports & Energy, and Pharma & Health.

WeBuild-AI is a trusted AI consulting partner focused on building production-grade AI solutions for global enterprises. They offer end-to-end services including strategy and roadmap development, custom AI solution design and deployment, and AI agents for automation. The company also specializes in architecting AI-ready data and infrastructure, AI-native engineering, and AI operating model design. WeBuild-AI helps establish responsible AI frameworks for governance and risk management, ensuring ethical use and regulatory compliance. Their AI Launchpad, the Pathway Platform, delivers proof-of-value capabilities rapidly, with most clients seeing measurable ROI within 10 weeks of pilot deployment. They integrate securely with existing systems using APIs and custom middleware.

FATE (Federated AI Technology Enabler) is an industrial-grade open-source framework designed for federated learning, hosted by the Linux Foundation. It facilitates secure data collaboration and privacy-preserving machine learning for enterprises and institutions. The framework implements secure computation protocols based on homomorphic encryption and multi-party computation (MPC). FATE supports various federated learning scenarios and provides a host of algorithms, including logistic regression, tree-based algorithms, deep learning, and transfer learning. It can be deployed on single or multiple nodes, with options for PyPI, Docker images, or CLI-based cluster deployment. The project also includes related tools like KubeFATE for cloud-native operations, FATE-Flow for task scheduling, and FATE-Board for visualization.

NeuroGaint Systems (NGS) delivers comprehensive digital transformation services, specializing in AI, automation, and cloud solutions for enterprises. With over 25 years of expertise as an IBM Business Partner, NGS offers deep capabilities in IBM watsonx, FileNet, Datacap, and CP4BA. Their services include AI-powered data and analytics, application development, cloud services, and DevOps containerization. NGS has developed NeuroLC, an AI-powered Trade Finance solution for Letter of Credit management, and serves diverse industries including finance, retail, manufacturing, and technology. They aim to empower businesses with scalable, secure, and tailored software solutions to drive growth and business transformation.

Nowigence provides comprehensive AI data analytics and business intelligence solutions, encompassing both software and hardware. Its no-code platforms, such as Nowg AI, enable users to build custom apps and automate workflows, while ResearchWork AI extracts and classifies insights from multiple documents. Tagion AI offers rapid data labeling and annotation services with a network of over 200,000 labelers. The Agri AI platform optimizes farm yields using AI and IoT. Nowigence also offers an AI Marketplace, consulting services, and robust cloud infrastructure for reliable and secure AI application deployment, focusing on enhancing human capabilities through AI-assisted labeling and data engineering automation.

Mode Maison is an independent research lab dedicated to exploring the intersection of physics-based simulation, artificial intelligence, material science, design, and retail. The company is actively developing multimodal Large World Models (LWMs) designed to simulate and generate diverse experiences. Their work aims to introduce a new dimension of intelligence, creativity, and reality, leveraging AI to push boundaries in these fields. While specific features are not detailed, the focus is on advanced AI research and development, suggesting a highly technical and innovative approach to complex problems in scientific computing and related domains.

Sensory, Inc. specializes in developing high-accuracy, low-power on-device AI solutions for voice, sound, and biometrics. Their technology is integrated into over 3 billion devices, offering features like wake word detection, speech-to-text conversion, sound identification, and biometric verification. A key differentiator is its edge AI capability, meaning all processing occurs directly on the device, ensuring enhanced privacy, faster response times, and reduced power consumption without relying on cloud infrastructure. This approach makes Sensory's solutions ideal for applications requiring real-time, secure, and offline functionality across various industries, including automotive, consumer electronics, mobile, healthcare, and retail. They offer a suite of products such as Smart Wake Word, Speech-to-Text, SoundID, and Face Verification, all designed for embedded performance.

FindErnest offers comprehensive technology consulting and digital transformation services designed to empower business growth. They provide tailored strategies and global insights across various domains, including Artificial Intelligence, Cloud Engineering, Software Development, Cybersecurity, and Managed IT Services. FindErnest focuses on delivering impactful results by blending advanced technology with transformative growth strategies, aiming to boost customer satisfaction, optimize operations, and provide insightful data. Their services are built for agility, ensuring quick ROI, and they hold certifications like CMMI Maturity Level 5, ISO 9001:2015, and ISO 27001:2022, demonstrating their commitment to quality and security.

Fifty2 Global Solutions provides comprehensive digital transformation services, leveraging strategy, design, and technology to deliver world-class software and digital solutions. Their expertise spans custom development, artificial intelligence (AI) integration, and robust cybersecurity measures. They partner with businesses to navigate the complexities of digital evolution, ensuring tailored solutions that drive efficiency and innovation. Fifty2 focuses on delivering impactful results through a holistic approach, from initial strategic planning to the final implementation of advanced digital systems. Their services are designed to help organizations achieve their digital goals and maintain a competitive edge in the evolving technological landscape.

AlpacaFarm is a simulation framework designed for research and development in Reinforcement Learning from Human Feedback (RLHF) and related methods. It significantly reduces the cost and complexity associated with developing RLHF techniques by eliminating the need for extensive human data collection. The framework offers low-cost simulation of pairwise feedback using API models like GPT-4, automated evaluations for method development, and validated reference implementations of baseline methods such as PPO and best-of-n. This open-source tool promotes accessible research on instruction following and alignment, making it easier for researchers to experiment and iterate on new RLHF approaches.

AI-Research-SKILLs is a comprehensive open-source library designed to empower AI agents to autonomously conduct AI research. It offers 87 specialized skills across 22 categories, covering the entire research lifecycle from literature surveys and idea generation to experiment execution and paper writing. The library includes both research orchestration layers, such as autoresearch and ideation, and engineering skills for training, evaluation, and deployment. It aims to reduce the time AI researchers spend on debugging infrastructure, allowing them to focus on hypothesis testing. The skills are production-ready, with documentation sourced from official repositories and real-world workflows, and can be installed via an interactive installer or directly through the Claude Code CLI.

albert_pytorch offers a PyTorch implementation of the ALBERT (A Lite Bert For Self-Supervised Learning Language Representations) model. This open-source repository provides the necessary code and pre-trained English models for researchers and developers working with natural language processing. Users can download various versions of pre-trained ALBERT models (v1 and v2, including base, large, xlarge, and xxlarge) and fine-tune them for specific tasks. The repository also includes scripts for converting TensorFlow checkpoints to PyTorch, preparing language model data, and running classifiers on benchmarks like GLUE. It supports dependencies such as PyTorch, CUDA, and scikit-learn, making it a valuable resource for those looking to implement or experiment with ALBERT.

AINE AI is a deep tech company specializing in AI, machine learning, and blockchain solutions for businesses. They offer a three-pronged approach: proprietary products like AComm AI for intelligent communication and ACash for digital payments; custom development services for bespoke technology products, including AI/ML model development and blockchain solutions; and staff augmentation to provide skilled AI engineers, data scientists, and blockchain developers on demand. Their core capabilities include advanced data mining, machine learning algorithms, intelligent decision systems, and decentralized blockchain solutions, all exposed via simple RESTful APIs to facilitate integration and accelerate digital transformation.

Energy Robotics offers an AI software platform designed for autonomous inspection robots and drones, catering to large-scale industrial facilities. The platform is hardware-agnostic, supporting various robots and drones like ExR-2, RB-WATCHER, Capra Scout, Inspector, Mavic 3T, and Spot. It provides AI-driven insights for predictive maintenance, transforming raw inspection data into actionable business intelligence and delivering timely alerts. Key features include AI data processing, API interface, and robot fleet management. The software equips robots with AI skills such as gauge reading, valve detection, audio analysis, people detection, missing objects detection, and fence detection, helping to relieve humans from dull and dangerous tasks and reduce operational costs. It is particularly beneficial for industries like Oil & Gas, Chemical, and Power & Utilities, enabling 24/7 automated inspections.