AI Agents & Automation

SideChat is a ChatGPT-powered AI assistant designed for seamless integration into your browser workflow. It allows users to access AI capabilities directly from any browser tab, providing instant assistance without switching applications. The tool leverages the latest ChatGPT model to offer real-time data and responses, ensuring up-to-date information. SideChat also features specialized assistants tailored for various tasks, enhancing productivity across different use cases. It prioritizes user privacy, making it a secure option for integrating AI into daily browsing activities. This makes it an efficient solution for individuals seeking to streamline their online tasks with AI support.

EverMemOS is a memory operating system designed to equip AI agents with persistent, proactive, and self-evolving memory. It addresses the limitations of stateless LLMs by enabling them to maintain context across days, sessions, and platforms, effectively turning them into intelligent agents that can truly remember. The tool offers multimodal retrieval and ingestion (mRAG), allowing it to parse and store various data types like PDFs, images, spreadsheets, and URLs through a single API. EverMemOS records agent trajectories as 'Cases,' distills patterns into reusable 'Skills,' and facilitates learning from past experiences. It also features a 'Memory Bank' interface for transparent management of user, group, and agent memories, including the ability to inspect and edit generated Skills. EverMemOS is available as a cloud service or an open-source solution.

Humains is a cutting-edge platform designed for generating and managing proactive autonomous AI agents for enterprise-level tasks. Powered by its Cognitive OS, an architecture inspired by the human brain, Humains agents can engage in millions of conversations, perform tasks like lead qualification, meeting scheduling, debt collection, and in-car assistance. The platform emphasizes hallucination-resistant AI, continuous optimization through patent-pending training methods, and comprehensive control for businesses to manage agent performance and alignment. Humains offers an end-to-end solution for proactive customer care, integrating with existing APIs to trigger events without complex coding, and features advanced voice recognition with emotional intelligence for natural interactions.

Gatik AI Inc. specializes in AI-powered autonomous trucking solutions for regional logistics networks, focusing on the middle-mile segment of supply chains. Their core technology, Gatik Driver™, is an AI system purpose-built for autonomous trucks, designed to handle real-world complexities like highway merges, dense urban traffic, and dock-to-dock precision. This system, integrated with production-grade hardware, allows for safe, scalable, and interpretable autonomous operations. Gatik's technology enhances the reliability and cost efficiency of B2B supply chains, proven in driverless commercial operations for Fortune 50 customers. They partner with industry leaders and operate in multiple markets, offering an Autonomous Transportation as a Service (ATaaS) model.

notebooklm-mcp is a server designed to bridge the gap between AI agents like Claude Code, Codex, and Cursor, and Google's NotebookLM. It allows these agents to directly query NotebookLM for information, ensuring zero-hallucination answers grounded in your own documentation. The tool addresses common problems with traditional AI research, such as massive token consumption, inaccurate retrieval, and hallucinations, by leveraging NotebookLM's pre-processed knowledge base. It supports persistent authentication, library management with tags and descriptions, and cross-client sharing, enabling deep, iterative research where agents automatically ask follow-up questions to build complete understanding before generating code or responses. This eliminates the need for manual copy-pasting and significantly improves the accuracy and efficiency of AI-driven research.

OpenCog is an open-source framework designed for integrated Artificial Intelligence and Artificial General Intelligence (AGI). While the original repository is no longer actively maintained, it served as a comprehensive laboratory for exploring and integrating diverse AI algorithms and systems, with a particular focus on humanoid robotics. Key areas of development included natural language chat, common-sense reasoning, various learning algorithms, and motor control for robots. The project aimed to develop AGI systems, and the framework compiled decades of work from numerous researchers. Many core components have since been split into separate, actively maintained repositories like atomspace, cogserver, and lg-atomese, making the original repository largely obsolete but still a valuable historical reference.

OpenContracts is an open-source platform designed to facilitate collaboration between humans and AI agents in building and managing knowledge bases from documents. It addresses the challenge of knowledge trapped in contracts, regulations, and research papers by providing tools for document annotation, version control, and semantic search. The platform allows teams to define custom label schemas, annotate documents with precise selections, and map relationships between concepts. It features configurable AI agents that can search documents, query annotations, and participate in discussions, all grounded in the structured knowledge created by the team. OpenContracts supports PDF and text-based formats, with upcoming DOCX viewing, and includes a modular processing pipeline for custom parsers, embedders, and thumbnail generators. It emphasizes human annotation as the ground truth, with AI building upon this curated data rather than replacing it.

OpenFaceIOS is an iOS project derived from OpenFace, a leading open-source tool for advanced facial analysis. This project specifically targets the iOS platform, streamlining the original OpenFace capabilities by removing dependencies like Dlib, TBB, and Boost, and exclusively utilizing OpenCV for its operations. Key functionalities include precise facial landmark detection, accurate head pose estimation, and sophisticated eye-gaze tracking. Additionally, OpenFaceIOS incorporates a unique face-swapping feature, allowing users to swap faces between individuals or apply model faces to real-time input. It's an ideal solution for developers and researchers looking to integrate robust facial analysis and manipulation into iOS applications.

The openai-cua-sample-app is a TypeScript sample application designed to teach users how to integrate and utilize OpenAI's Computer Using Agent (CUA) via its API across various computer environments. This repository includes a Next.js operator console (apps/demo-web) for managing runs, reviewing screenshots, events, and replay artifacts, alongside a Fastify runner (apps/runner) that handles mutable workspaces, browser sessions, and SSE. It showcases how to switch between code mode and native computer mode, define scenario manifests, and build an operator-facing console. The sample app is ideal for developers looking to understand and implement browser-focused AI automation with GPT-5.4, providing a clear example of integrating the Responses API.

OpenViking is an open-source context database tailored for AI Agents, addressing common challenges in agent development such as fragmented context, surging demand, poor retrieval, and unobservable context. It innovatively adopts a "file system paradigm" to unify the structured organization of memories, resources, and skills, moving away from traditional RAG's fragmented vector storage. Key features include tiered context loading to reduce token consumption, directory recursive retrieval for improved accuracy, visualized retrieval trajectories for observable context, and automatic session management for context self-iteration. OpenViking supports various VLM and embedding model providers, including Volcengine, OpenAI, and local models via Ollama, making it a versatile solution for developers building intelligent AI agents.

PaddlePaddle is China's first independent R&D deep learning platform, open-sourced since 2016. It offers an industrial-grade machine learning framework with advanced technologies and rich features, covering core deep learning frameworks, basic model libraries, end-to-end development kits, tools, components, and service platforms. Originating from industrial practices, PaddlePaddle is dedicated to industrialization and has been widely adopted across various sectors. Key features include unified dynamic/static graphs with automatic parallelism, integrated training and inference for large models, high-order differentiation for scientific computing, and a neural network compiler. It also provides heterogeneous multi-chip adaptation through standardized interfaces, supporting a pluggable architecture for diverse hardware.

PentestAgent is an open-source AI agent framework designed for black-box security testing, supporting bug bounty, red-team, and penetration testing workflows. It requires Python 3.10+ and an API key for OpenAI, Anthropic, or other LiteLLM-supported providers. The framework features multiple operational modes, including single-shot instruction execution, autonomous task execution, and a multi-agent crew mode where an orchestrator spawns specialized workers. It also offers an interactive chat mode for guided pentesting. PentestAgent supports running tools in Docker containers for isolation and pre-installed pentesting utilities, and includes prebuilt attack playbooks for structured security assessments. It integrates with the Model Context Protocol (MCP) for extensibility, allowing consumption of external MCP servers and exposing itself as an MCP server for remote control.

Oxen is a lightning-fast, open-source data version control system specifically designed for large machine learning datasets. It supports both structured and unstructured data, including images, audio, video, text, and tabular data like parquet files. The system aims to make data versioning as straightforward as code versioning, mirroring the interface of Git but optimized for the unique challenges of large datasets. Oxen can efficiently index and sync hundreds of thousands of files in seconds, scaling to terabytes of data. It offers a command-line interface, along with bindings for Rust and Python, and HTTP interfaces for easy integration into existing workflows. Key features include efficient indexing, handling of large files, native DataFrame processing, change tracking, and collaboration capabilities through an Oxen server.

agent-sandbox is a Kubernetes-native project developing a Sandbox Custom Resource Definition (CRD) and controller designed for easy management of isolated, stateful, singleton workloads. It's particularly well-suited for use cases like AI agent runtimes, development environments, and persistent single-container sessions for tools like Jupyter Notebooks. The core Sandbox CRD offers a declarative API for managing a single, stateful pod with stable identity and persistent storage, addressing limitations of standard Kubernetes Deployments and StatefulSets for these specific needs. Key features include stable identity, persistent storage, and comprehensive lifecycle management. Extensions like SandboxTemplate, SandboxClaim, and SandboxWarmPool further enhance its capabilities by providing reusable templates, user-friendly provisioning, and pre-warmed pools for rapid allocation.

Powered By Agency specializes in developing custom AI agents designed to automate customer interactions for small and medium-sized businesses. These agents can handle communications across various channels, including voice, email, and SMS, streamlining operations and enhancing customer engagement. The service focuses on creating personalized AI solutions that are tailored to the specific needs of each business, aiming to boost efficiency and accelerate growth. By deploying these advanced AI agents, SMBs can automate routine tasks, improve response times, and free up human resources to focus on more complex strategic initiatives. This digital marketing agency emphasizes transforming business operations through intelligent automation.

Paper_Reading_List is a comprehensive resource offering a curated collection of recommended papers in the fields of Computer Vision and Pattern Recognition (cs.CV), Artificial Intelligence (cs.AI), and Learning (cs.LG). This open-source GitHub repository aims to provide a structured reading roadmap for individuals new to deep learning, helping them navigate the vast amount of research available. The list is organized by topics such as Salient Object Detection, Object Detection, Semantic Segmentation, and Visual Question Answering, making it easier for users to find relevant papers. It also includes links to other valuable deep learning paper collections and notes, enhancing its utility as a central hub for research.

papers-for-molecular-design-using-DL is an open-source repository offering a curated list of academic papers focused on molecular and material design utilizing generative AI and deep learning techniques. This resource is particularly valuable for researchers and data scientists working in drug discovery, materials science, and computational chemistry. The list covers various aspects, including unified deep molecular generative models, generative AI for scientific discovery, and deep learning-based drug design. It also includes sections on molecular optimization, material design, and molecular conformation generation, providing a structured overview of the latest advancements and research trends in these fields. The repository is continuously updated, ensuring access to current literature and methodologies.

Sanctuary AI is dedicated to creating and deploying millions of industrial-grade humanoid robots to tackle growing labor challenges globally. Their Phoenix robots are general-purpose, AI-powered humanoids built for autonomous work in sectors like automotive, manufacturing, and logistics. These robots focus on dexterity, tactile feedback, and fine manipulation, featuring a strong torso for industrial tasks. Sanctuary's embodied AI systems mimic human movement and cognitive processes, enabling the AI to learn and interact with the world similarly to humans. The company aims to provide solutions for jobs that are often considered dull, dirty, or dangerous, thereby securing the continued flourishing of civilization amidst declining populations and labor shortages.

AI2-THOR is an open-source platform developed by the Allen Institute for AI (AI2) designed for Visual AI research. It offers a near photo-realistic and interactable framework for embodied AI agents, supporting research in common sense reasoning. The platform includes various environments such as iTHOR for high-level interaction, ManipulaTHOR for visual object manipulation with robotic arms, and RoboTHOR for Sim2Real research with simulated and physical world counterparts. It features over 200 custom-built scenes, 2600+ heavily annotated household objects with realistic physics, and multiple agent types including multi-agent support, LoCoBot, and Kinova 3 inspired robotic manipulation agents. AI2-THOR also provides 200+ actions for interaction and navigation tasks, first-class support for various image modalities (RGB, instance/semantic segmentation, depth, normals), and extensive metadata for complex reward functions.

prime-diloco, previously known as ZeroBand, is an open-source framework designed for efficient and globally distributed training of AI models across the internet. It introduces an ElasticDeviceMesh for fault-tolerant training, allowing dynamic resizing of global process groups when nodes join or leave, unlike standard DeviceMesh implementations. The framework also features asynchronous distributed checkpointing to minimize blocking time during training, live checkpoint recovery for nodes joining mid-training, and a custom Int8 All-Reduce kernel for significant payload size reduction without impacting loss curves. Additionally, it maximizes bandwidth utilization through sharding and VPN technology, and implements PyTorch FSDP2 / DTensor ZeRO-3 for memory optimization and CPU off-loading for its optimizer.

The Inverse Scaling Prize was a contest designed to find tasks where larger language models (LLMs) exhibit inverse scaling, meaning their performance degrades as they become larger and more capable at general language modeling. The initiative aimed to highlight potential failure modes in the current paradigm of LLM pretraining and scaling, emphasizing the importance of understanding these issues for the safe and responsible use of AI. Participants were challenged to submit tasks demonstrating this counter-intuitive behavior, with significant prize money awarded to winning entries. The contest concluded with two rounds of submissions, and the organizers plan to publish a paper surveying the submitted tasks, inviting authors of winning and accepted tasks to co-author. The project also released data for all winning tasks, contributing to ongoing research in LLM behavior.

Poplar is a web-based annotation tool specifically designed for natural language processing (NLP) needs. It is inspired by the well-known brat rapid annotation tool, providing a platform for efficient text annotation. As an open-source project, it allows developers and researchers to integrate and customize it for their specific requirements. The tool is the new version of synyi-annotation-tool, indicating continuous development and improvement. While it is noted as not yet production-ready, it offers a robust foundation for NLP annotation tasks and encourages community contributions for further enhancement.

Pipcook is an open-source JavaScript application framework designed for machine learning and its engineering, specifically targeting web developers. It aims to empower JavaScript engineers to leverage machine learning capabilities without needing extensive prerequisites in the field. The platform focuses on bridging machine learning with front-end interaction, offering a modular and flexible architecture. Key components include a machine learning pipeline framework, management tools, and a JavaScript runtime for machine learning. Pipcook also features a bridge to Python, allowing JavaScript developers to access Python packages like NumPy, scikit-learn, and TensorFlow within a Node.js environment. This makes it easier to train models, serve them, and optimize their performance for various applications like image classification and object detection.

auto-gpt-web is an open-source AI agent that empowers users to define initial roles and goals for an AI buddy, which then operates autonomously to achieve these objectives. Inspired by Auto-GPT, this tool features internet access for comprehensive searches and information gathering. A key differentiator is its local storage capability, saving AI definitions, chat history, and credentials directly within the user's browser, ensuring privacy and control. It also includes long-term memory based on a browser-based vector database and an Electron application for conducting search operations, bypassing typical API limitations. Users need an OpenAI API Key and Google Search API Key with a Custom Search Engine ID to utilize its full capabilities.