AI Agents & Automation

Beyond Vision is a leading drone manufacturer focused on integrating AI algorithms and cloud-based management platforms into their UAVs. Developed to NATO standards, their drones offer fully autonomous flight, obstacle avoidance, detection, and tracking capabilities. The beXStream platform, a cloud-based asset management system, uses AI to integrate various UAV systems and conventional automation, optimizing asset management and driving operational efficiency. Beyond Vision's commitment to R&D, in collaboration with scientific institutions, aims to expand possibilities in the UAV industry. They offer a range of products including hexacopters, quadcopters, fixed-wing drones, remote control software, and RTK base stations, serving defense, agriculture, infrastructure, and safety sectors.

Redtail is an open-source project by NVIDIA-AI-IOT, offering perception and AI components for autonomous mobile robotics. It enables visual navigation for drones and ground vehicles in complex, unstructured environments like forest trails. The project includes deep neural networks (DNNs), computer vision code, control code, and hardware instructions. Users can train DNNs from scratch using public data, and pre-trained models are also available. Redtail supports NVIDIA's Jetson embedded platform and features modules like TrailNet DNN for visual navigation and Stereo DNN for depth estimation from stereo cameras. While no longer actively developed by NVIDIA, the community continues to provide alternative implementations.

DigitalNet.ai offers a comprehensive suite of enterprise AI solutions designed to drive business transformation. Their core offering, JanusAI, is a real human AI platform built for business, providing agentic AI solutions. Beyond agentic AI, DigitalNet.ai specializes in data analytics to transform data, advanced cybersecurity defenses, and digital transformation services to modernize enterprises. The platform boasts over 1,500 customers served, leading to 60%+ faster decision-making, and utilizes over 2,000 pre-built cognitive agents and 70+ advanced AI models. DigitalNet.ai is compliant with SOC2, HIPAA, GDPR, and FedRAMP standards, ensuring secure and responsible AI deployment.

Awesome-LLM4IE-Papers is a valuable open-source resource hosted on GitHub, dedicated to compiling academic papers focused on generative Information Extraction (IE) using Large Language Models (LLMs). The repository is organized based on a comprehensive survey titled "Large Language Models for Generative Information Extraction: A Survey," which also serves as a guide to the paper organization. It categorizes papers by specific IE tasks such as Named Entity Recognition, Relation Extraction, and Event Extraction, as well as by techniques like Supervised Fine-tuning, Few-shot learning, and Prompt Design. Researchers and practitioners can find recently updated papers, datasets with download links, and contribute suggestions or corrections via email. This resource is essential for staying current with advancements in LLM-driven information extraction.

Glambase is a cutting-edge platform designed for revolutionizing digital influence by enabling users to create, customize, and monetize their own AI-driven virtual influencers. The platform offers an easy-to-use character face generator, allowing for the design of unique virtual personas with distinct charm and energy. Users can explore a variety of pre-existing AI characters or build their own, supporting their favorites and chatting with them. Glambase aims to help users generate revenue through their AI influencers, saving time by managing interactions seamlessly. It's presented as a powerful tool for creating and monetizing AI influencers with an extremely easy process.

Vyomastra Technologies is at the forefront of drone innovation, manufacturing advanced Unmanned Aerial Vehicles (UAVs) designed for diverse industrial applications. Their product line includes the Voyager 5G, which ensures global connectivity and integrated obstacle avoidance, capable of transporting payloads up to 2.5 KG with over 30 minutes of flight duration and real-time streaming. The Bhoomi drone offers PPK/RTK compatibility, Lidar sensors for terrain following, and autonomous functionality, covering large areas efficiently with over 40 minutes of flight time and high-resolution imaging. Vyomastra also provides Ground Control Station software for land surveying, inspections, and surveillance, featuring real-time AI inference and analytics integrated with live video streaming across all devices.

BrainBox AI is an advanced AI building management platform designed to autonomously optimize HVAC systems and operational processes within commercial buildings. It leverages both autonomous and generative AI technologies to significantly cut energy consumption, reduce greenhouse gas emissions, and enhance overall building efficiency. The platform includes AI Control for HVAC optimization and a Cloud Building Management System for comprehensive portfolio control. BrainBox AI is suitable for various building types, from retail and distribution centers to corporate offices and academic institutions, helping organizations achieve measurable reductions in energy use and carbon footprint while improving operational performance and streamlining processes. It offers solutions like ARIA, an AI building engineer, and provides resources such as case studies, white papers, and tools for achieving net-zero buildings.

agentic-cursorrules provides a practical approach to managing multiple AI agents within the Cursor code editor by partitioning large codebases into domain-specific contexts. This prevents agent conflicts by giving each agent explicit file-tree boundaries. The tool generates isolated markdown rule files, such as `@agent_backend_api.md`, which describe these boundaries, relevant paths, and guardrails for specific domains. Users can reference these files in their IDEs (like Cursor or Windsurf) to ensure AI helpers operate within their defined lanes. It supports auto-detection of logical domains and interactive configuration, or manual boundary definition via a `config.yaml` file. The tool helps keep conversations focused, diffs local, and reduces coordination overhead by preventing agents from trying to understand an entire repository at once.

AgentLab is an open-source framework developed by ServiceNow for the creation, evaluation, and benchmarking of web agents. It provides a scalable and reproducible environment for researchers and developers to work with various benchmarks supported by BrowserGym. Key features include easy large-scale parallel agent experiments using Ray, a unified LLM API for OpenRouter, OpenAI, Azure, and self-hosted models, and support for running benchmarks like WebArena and WorkArena. The framework also offers various reproducibility features, a unified leaderboard, and tools like AgentXray for visualizing experiment results through a Gradio interface. It is designed to accelerate web agent research by providing robust building blocks and evaluation capabilities.

Agent-R1 is an open-source framework specifically designed for training powerful language agents using end-to-end reinforcement learning. It excels in multi-step agent tasks where models interact with environments and tools over multiple rounds, rather than producing a single final answer. The core innovation is its Step-level MDP, treating each interaction step as a proper RL transition with environment-defined states, LLM actions, and next observations. This approach addresses common issues like retokenization drift and rigid token-only trajectory construction found in traditional text-based pipelines. With Agent-R1, users can build custom agent workflows, define interactive environments and tools, and train multi-step agents within a unified RL pipeline, making it ideal for complex agent development.

BMC Helix delivers cognitive service management that is fast, accurate, and cost-effective across multi-cloud, multi-device, and multi-channel environments. It provides a fleet of AI agents to reduce manual effort and optimize resources. Key offerings include AI-driven IT service management (ITSM), an AI-powered virtual assistant for conversational search, and exceptional customer service management. For employee engagement, it offers personalized experiences, intuitive self-service, and modern case management. Additionally, BMC Helix provides AIOps and Observability solutions for discovery, monitoring, remediation, and optimization, alongside SecOps and Compliance features to reduce risk and streamline regulatory responses.

OpenPipe is an AI tool designed to optimize AI models and enhance developer efficiency, particularly for those working with Reinforcement Learning (RL) for agents. It focuses on improving AI workflows and streamlining the development process through seamless integration capabilities. The platform aims to provide robust solutions for building and managing AI agents, ensuring high-quality output and efficient operation. OpenPipe supports developers in creating more effective AI systems by offering tools that simplify complex RL implementations and integrate smoothly into existing development environments. This focus on optimization and integration makes it a valuable asset for developers looking to push the boundaries of AI agent performance.

Runbear is an AI agent platform designed to automate Slack inboxes for operations, customer success, and RevOps teams. It integrates with over 2,000 tools, including Notion, Linear, HubSpot, and Google Drive, to pull answers and take action directly within Slack, Microsoft Teams, and HubSpot. The platform acts like a teammate, understanding context, answering questions, routing inquiries, creating tickets, and updating CRMs without requiring users to leave their communication platforms. Runbear learns from team interactions and existing documentation, getting smarter with every conversation. It boasts a quick 10-minute, no-code setup, allowing users to connect tools and invite Runbear to channels easily. Security features include SOC 2 Type II compliance, SSO, RBAC, and data encryption, ensuring data safety and privacy.

AI Strategy Corporation, through its 9-HI platform, offers a patent-pending technology for Human-AI collaborative Decision Guidance. This platform is designed to enhance success in commercialization by integrating project management, innovation management, knowledge management, risk management, and state-of-the-art Human and AI collaboration. It aims to drive product or technology success, funded by entities like the US Air Force and OUSD Defense Production Act. 9-HI helps organizations make more informed decisions on technology investments, providing a standard process for technology acquisition and development. The platform utilizes AI Agents like Orion, Scorpius, Cassiopeia, Hercules, and Libra to assist with subject matter expertise, risk identification, success factors, solution options, and success evidence, feeding into a Success Probability Scoring Engine.

ZenCall.ai is an AI-powered tool designed to streamline multilingual call management, ensuring businesses never miss a customer call. It offers seamless app integration, allowing for enhanced functionality and improved communication workflows. The platform is ideal for automating customer support and personal assistance tasks, making it easier to handle diverse customer bases. By leveraging AI, ZenCall.ai aims to improve efficiency and consistency in customer interactions, providing a robust solution for businesses looking to optimize their call handling processes and maintain high service standards across different languages.

Alethea AI is a research and development studio at the forefront of Generative AI and Blockchain, pioneering Sovereign AI Agents. Their platform allows users to activate customized video agents for educational content, marketing, and community interaction. It also facilitates the creation of digital twins to extend online presence and engage followers. Furthermore, Alethea AI offers real-time video agents built for scale, combining their AI engine with a scalable intelligence framework to automate tasks and deliver reliable multimodal outputs for enterprises. The company is focused on creating autonomous, emotionally intelligent AI owned by users, not corporations, empowering humanity through decentralized AI.

Zamp provides advanced digital employees, specifically its AI agent named Pace, designed to handle operational work that humans shouldn't have to do. Pace learns from existing business processes, integrates with various systems, and uses feedback to automate manual tasks with high accuracy. It can extract documents, automate browser actions, make decisions, and communicate, getting up to speed within four weeks. Zamp emphasizes enterprise-grade security, offering end-to-end encryption, audit trails, and SOX compliance, making it suitable for leading public companies and financial institutions. The platform aims to free human teams to focus on strategic growth and innovation by handling 99% of operational work.

Skywork-OR1 (Open Reasoner 1) is an open-source project that unleashes the power of reinforcement learning for developing advanced math and code reasoners. It features a series of powerful models, including Skywork-OR1-7B and Skywork-OR1-32B, which are trained using large-scale rule-based reinforcement learning with carefully designed datasets and training recipes. The 32B model notably outperforms Deepseek-R1 and Qwen3-32B on math tasks (AIME24 and AIME25) and delivers comparable performance on coding tasks (LiveCodeBench). The project provides comprehensive resources, including model weights, training data, and code, to help researchers and developers advance the field of open reasoning models. It also offers detailed documentation for installation, training, and evaluation.

AutoChain is a robust framework designed for developers to build lightweight, extensible, and testable LLM agents. It addresses the challenges of customizing generative agents for specific purposes and the manual, repetitive nature of evaluating them. Inspired by LangChain and AutoGPT, AutoChain offers a simplified approach to agent development, allowing for rapid iteration through easy prompt updates and minimal abstraction layers. A key differentiator is its automated multi-turn conversation evaluation framework, which uses LLM-simulated test users to assess agent performance across various scenarios, ensuring comprehensive testing and reducing regression risks. It supports custom tools and OpenAI function calling, making it versatile for different generative agent applications.

AutoHedge is an enterprise-grade autonomous agent hedge fund platform designed to automate market analysis, risk management, and trade execution. It leverages swarm intelligence and specialized AI agents to perform continuous analysis, generate and validate trading theses, size risk, and execute orders across supported venues. The system is built for institutional reliability, featuring structured outputs, comprehensive logging, and a risk-first architecture. It currently supports full autonomous trading on Solana, with plans to expand to Coinbase and other exchanges. Its multi-agent architecture includes Director, Quant, Risk Management, and Execution Agents, ensuring a robust and scalable solution for financial professionals.

any-agent offers a unified interface for interacting with and assessing diverse AI agent frameworks. This tool is designed to streamline the development and deployment of AI-powered agents by providing functionalities for tracing agent activities and serving them efficiently. It supports a range of existing frameworks and is actively seeking contributions for new ones, making it a flexible solution for AI researchers and developers. The platform emphasizes ease of integration with different models and includes practical examples and cookbooks to help users quickly get started with building and evaluating agents, including multi-agent systems.

Atomic Agents offers a lightweight and modular framework for constructing AI agent pipelines and applications. Designed with atomicity in mind, it allows developers to build AI systems using single-purpose, reusable, and composable components, ensuring predictable and consistent outputs. The framework is built on Instructor and Pydantic, enabling the application of standard software engineering practices to AI development. Key features include defining clear input and output schemas, dynamic context injection via Context Providers, and easy chaining of agents and tools. This approach provides enhanced control and extensibility, making it ideal for creating robust and maintainable AI applications.

Auto-GPT-Plugin-Template serves as a foundational resource for developers looking to create custom plugins for Auto-GPT. This template provides a structured environment, including necessary files and configurations, to streamline the plugin development process. It guides users on how to integrate their custom functionalities, manage dependencies, and enable their plugins within the Auto-GPT ecosystem. The template is specifically designed for plugins that operate externally to the core Auto-GPT codebase, offering clear instructions on installation and configuration. Developers can leverage this template to extend Auto-GPT's capabilities with new tools and features, ensuring compatibility and ease of integration.

Centific helps model labs and enterprises build, train, deploy, and govern intelligent systems by providing high-quality data, human expertise, and end-to-end platforms. The company focuses on generating, refining, and operationalizing real-world signals across language, vision, behavior, and expertise to enable AI systems to learn faster and perform better in production. Centific offers solutions for RL Environments-as-a-Service, Translation & Localization, Multilingual AI, Data Collection & Creation, RLHF & Preference Optimization, Supervised Fine Tuning, Model Safety & Evaluation, and Internationalization. Their platforms include Data Marketplace, Data Canvas, AI Data Foundry, and OneForma, designed to support continuous data loops for production AI.