AI Agents & Automation

Muks Robotics is India’s leading AI humanoid robotics company, specializing in building autonomous enterprise humanoids. Their 'Spaceo' line of humanoids, including M1, PRO, and PRIME models, are designed to address labor shortages and automate dangerous or repetitive tasks across various industries. The humanoids are powered by Fusion Max, a Vision-Audio-Language-Action AI model with 2 billion optimized parameters, trained on real-world data. Key features include autonomous task execution, advanced navigation, multimodal communication, human-centered modular design, and built-in safety systems, ensuring reliable performance in real-world environments. Muks Robotics aims to enable humanity’s future on Earth and beyond by providing intelligent, adaptable workforce solutions.

agenticSeek offers a 100% local alternative to cloud-based AI assistants, ensuring complete privacy by running entirely on your hardware. This voice-enabled AI assistant can autonomously browse the web, extract information, fill forms, and write/debug code in multiple languages like Python, C, Go, and Java. It intelligently selects the best agent for a given task and can break down complex projects into manageable steps. Designed for local reasoning models, agenticSeek keeps all data on your device, eliminating cloud dependency and monthly bills. It supports various local LLM providers like Ollama and LM Studio, and can also be configured to use API-based providers if local hardware is insufficient.

Agent-S is an open-source framework designed to enable autonomous interaction with computers through an Agent-Computer Interface. Its mission is to build intelligent GUI agents capable of learning from past experiences and performing complex tasks autonomously. The framework has seen significant advancements, with Agent S3 being the first to surpass human-level performance on OSWorld. It supports various models and providers, including Azure OpenAI, Anthropic, Gemini, Open Router, and vLLM inference, and can be configured with grounding models like UI-TARS-1.5-7B. Agent-S offers a local coding environment for tasks requiring code execution, such as data processing and system automation, providing flexibility for technical users.

Payman AI is an agentic AI platform designed for financial institutions to automate banking transactions. It enables the execution of payments, transfers, and account analysis through voice or text interfaces, leveraging a bank's existing infrastructure. The platform ensures full visibility, configurable controls, and complete audit trails for every transaction. Payman AI focuses on understanding customer intent beyond keywords, applying bank-specific rules automatically, and providing a seamless customer experience across various channels like voice, mobile, web, and SMS. It is built with robust safety and security features, including SOC 2 certification, compliance readiness (KYC, BSA, AML), and immutable audit trails, making it suitable for regulated environments. The system integrates quickly with core banking systems like FIS, Fiserv, Jack Henry, and digital platforms such as Narmi and Q2.

MetaBrain Labs Inc. offers a platform for developing sensor-integrated AI agents that connect people, systems, and live data to guide intelligent conversations. These agents are designed for real-time data ingestion from various sensors, including wearables, clinical devices, and IoT. The platform provides a blueprint for deploying sensor-aware AI agents, offering hands-on enablement and patent-backed rights to operate. It focuses on adaptive conversational intelligence, context-aware, task-driven dialogue, and flexible, scalable deployment options. Unlike generic language models, MetaBrain Labs' agents are task-oriented systems driven by live data and proprietary logic to solve real-world problems dynamically, ensuring ownership and customization for the user.

LangChain provides a comprehensive engineering platform and open-source frameworks designed for developers to build, test, and deploy reliable AI agents. The platform, LangSmith, offers robust tools for observability, allowing users to trace agent execution and understand complex interactions. It also includes evaluation capabilities to score and improve agent performance using real-world usage data and human feedback. For deployment, LangSmith supports shipping and scaling agents in production with features like memory, conversational threads, and durable checkpointing. Additionally, LangChain offers open-source frameworks like deepagents, langchain, and langgraph for building various types of agents, from quick-start prototypes to reliable production systems with low-level control.

Capably is an enterprise AI technology platform designed to unlock operational efficiency through intelligent automation solutions. It empowers organizations to multiply employee efficiency by automating bulk work and powering critical tasks with agentic AI. The platform offers a proprietary stack combining power, governance, safety, and adaptability, featuring an AI Platform for work automation, an APA Engine for process automation, an NLMP Interface for easy automation, and an AI Capability Library with ready-to-go, industry-specific AI capabilities. Capably provides a full partnership, from discovery to daily operations, taking ownership to deliver measurable results and deploying in weeks to prove value quickly. It supports various industries including Media & Advertising, Retail, FMCG, Healthcare, Real Estate, and Finance.

dm_control is Google DeepMind's comprehensive software stack designed for physics-based simulation and Reinforcement Learning (RL) environments, built upon the MuJoCo physics engine. It offers Python bindings to the MuJoCo engine, a suite of RL environments, and an interactive viewer for real-time interaction. The package also includes libraries for composing and modifying MuJoCo MJCF models in Python, defining rich RL environments from reusable components, and additional libraries for custom tasks like multi-agent soccer. This open-source tool is ideal for researchers and developers working on advanced AI and robotics applications, providing a robust infrastructure for developing and testing continuous control algorithms.

Minded, previously known as AgentsForce, is an innovative AI agent platform designed to empower users to build and deploy AI agents by simply recording their work. This approach eliminates the need for complex API integrations, allowing agents to operate like humans across existing tools and systems. The platform offers an intuitive drag-and-drop AI editor for agent creation and customization, alongside an AI Recorder that captures screen actions to train agents. Minded is built for regulated industries, providing full audit trails, SSO, and robust permission management, ensuring data security and compliance. It supports processing documents in any format with human-level accuracy and allows for management of AI agents using natural language.

GRU4Rec is the original Theano implementation of the algorithm described in the "Session-based Recommendations with Recurrent Neural Networks" paper (ICLR 2016) and its follow-up. This open-source tool is specifically optimized for fast execution on GPUs, capable of processing up to 1500 mini-batches per second on a GTX 1080Ti. While official PyTorch and TensorFlow reimplementations exist, this original Theano version is noted for being significantly faster. It provides functionalities for training, evaluating, and saving/loading GRU4Rec models, with detailed configuration options for GPU usage and hyperparameter tuning. The project emphasizes the importance of using this original implementation due to observed flaws and performance issues in third-party versions.

GPTQ-for-LLaMa offers a 4-bit quantization solution for LLaMA models, leveraging the GPTQ one-shot weight quantization method. This tool is specifically optimized for Linux operating systems and recommends the use of AutoGPTQ for enhanced performance and broader compatibility. While it can be applied universally, it may not be the fastest quantization method available. The project provides detailed benchmarks comparing its performance against FP16, RTN, and bitsandbytes for various LLaMA model sizes (7B, 13B, 33B, 65B) across different bit and group-size configurations, highlighting memory usage and checkpoint sizes. Installation instructions are provided for Conda and pip, along with dependencies and examples for language generation and model inference.

KawaiiGPT is an open-source AI agent tool available on GitHub, designed for educational purposes and experimentation. It features a reverse-engineered LLM API wrapper, building upon original agents from the Pollinations GitHub repository. The tool utilizes a server to integrate and serve various obtainable LLMs such as DeepSeek, Gemini, or Kimi-K2. It emphasizes that it uses prepared models with prompt injection for jailbreaking, rather than fine-tuned models. The project was created for fun and learning, with a disclaimer that users are responsible for their actions. The developer also addresses concerns about obfuscation, stating it's to prevent unauthorized re-selling and renaming of the tool.

LeRobot, developed by Hugging Face, aims to democratize AI for robotics by offering a comprehensive open-source framework for end-to-end learning. It features a hardware-agnostic, Python-native interface for standardized control across diverse robotic platforms, from low-cost arms to humanoids. The tool introduces a standardized, scalable LeRobotDataset format (Parquet + MP4/images) hosted on the Hugging Face Hub, facilitating efficient storage, streaming, and visualization of large robotic datasets. LeRobot also implements state-of-the-art policies in pure PyTorch for Imitation Learning, Reinforcement Learning, and Vision-Language-Action models, with tools for instrumenting and inspecting the training process. It supports evaluation in simulation or on real hardware using a unified script, including standard benchmarks like LIBERO and MetaWorld.

libonnx is a lightweight, portable pure C99 ONNX inference engine specifically designed for embedded devices. It offers hardware acceleration support, making it ideal for deploying AI models on systems with limited resources. The library's .c and .h files can be easily integrated into any project. Users can allocate an ONNX context, load models from files, search for input and output tensors, run the inference engine, and then free the context. It supports ONNX version v1.17.0 with opset 24 and includes tools for converting ONNX models into C arrays for embedded use. The project provides clear compilation instructions, cross-compilation examples, and methods for running tests and examples, such as MNIST handwritten digit prediction.

Autonodyne LLC is a Boston-based autonomous software company specializing in AI and smart software for command and control (C2) of unmanned vehicles. Their solutions support multi-vehicle swarming, allowing teams of unmanned vehicles to perform missions across air, sea, and land. The software also facilitates Manned and Unmanned Vehicle Teaming (MUM-T), reducing the cognitive burden on human operators and enhancing mission capabilities. Autonodyne's technology is hardware-agnostic, supporting over 60 makes/models of unmanned platforms, 15 communication protocols, and 16 datalink radios. They offer a library of autonomy behaviors for mission-specific maneuvers and provide services for counter-UAS and red-teaming.

BigPanda offers an agentic ITOps platform that leverages AI to automate IT detection, triage, and resolution processes. This platform is designed to enhance operational efficiency, reduce downtime, and lower costs for IT teams. Key features include AI Incident Prevention, AI Detection & Response, an L1 Agent, and an AI Incident Assistant. It also utilizes an IT Knowledge Graph to unify tribal knowledge, providing AI with context for reasoning. BigPanda aims to help enterprises prevent incidents, accelerate change approvals, and resolve issues faster, integrating with existing ITSM tools like Jira and ServiceNow to supercharge their value.

mqttclient is a robust, high-performance, and cross-platform MQTT client developed based on the socket API. It is designed for various environments, including embedded devices (FreeRTOS, LiteOS, RT-Thread, TencentOS tiny), Linux, Windows, and Mac. The client boasts extremely high stability, handling reconnections, packet loss, and retransmissions according to MQTT protocol standards. It is lightweight, consuming minimal resources, with the entire project code using less than 15KB of RAM without mbedtls. mqttclient supports mbedtls encrypted transmission for secure communication, offers a very simple API interface, and includes an online code generation tool. It also features automatic re-subscription of topics, support for theme wildcards, and a layered design for improved performance and reduced coupling.

MVSNeRF is a novel neural rendering approach presented at ICCV 2021, designed for efficiently reconstructing geometric and neural radiance fields to enable advanced view synthesis. This tool is implemented in PyTorch Lightning and facilitates fast per-scene reconstruction, especially when dense images are available for fine-tuning. It supports training on various datasets including DTU, Blender (Realistic Synthetic), LLFF (Real Forward-Facing), and custom data. Users can train models, fine-tune them for specific scenes, and render free-viewpoint videos. The repository provides detailed installation instructions, training commands, and evaluation scripts, making it a valuable resource for researchers and developers in 3D reconstruction and neural rendering.

NeuPAN (Neural Proximal Alternating-minimization Network) is an end-to-end, real-time, and map-free robot motion planner designed for direct point robot navigation. It integrates learning-based and optimization-based techniques to map obstacle points directly to control actions, ensuring high accuracy and safety in cluttered and unknown environments. Unlike traditional modular planners, NeuPAN avoids error propagation by eliminating middle modules and requires minimal training data, often just random points within a range. It boasts fast training times, typically 1-2 hours on a CPU for new robot geometries, and can be deployed without retraining for various environments. NeuPAN also provides a ROS wrapper for integration and supports DUNE model training for specific robot geometries.

OpenCat-Quadruped-Robot is an open-source framework designed for building and programming quadruped robots, inspired by Boston Dynamics' Spot. Developed by Petoi, it powers their Bittle robot dog and Nybble robot cat platforms. The framework simplifies complex tasks like gait coordination, servo control, and IMU integration, allowing users to focus on higher-level applications. It supports multiple languages including C/C++, Python, and block-based coding, and is compatible with Arduino and Raspberry Pi. OpenCat is utilized in K-12 schools, university research labs, and maker spaces for STEM education, IoT robotics, AI-enhanced applications, and DIY robotics kit development. It also supports sensor integration, simulation-to-real-world experiments, and is ROS compatible for advanced applications like SLAM and navigation.

Openclaw-skills, hosted on GitHub as BankrBot/skills, is a comprehensive open-source library designed to equip builders with plug-and-play tools for creating more powerful AI agents. It offers a diverse set of skills covering various domains, including on-chain financial operations like token launching, scam analysis, and liquidity management, as well as social media automation for platforms like Twitter/X and Farcaster. The library also includes tools for agent identity management, decentralized task marketplaces, and privacy-preserving transactions. Developers can contribute new skills via pull requests, expanding the capabilities of AI agents in finance, social interaction, and other areas. The project emphasizes modularity and ease of integration for agent builders.

PaddleSpeech is an open-source, easy-to-use speech toolkit built on the PaddlePaddle platform, designed for a variety of critical tasks in speech and audio. It features state-of-the-art and influential models, including self-supervised learning, streaming Automatic Speech Recognition (ASR) with punctuation, and streaming Text-to-Speech (TTS) with a robust text frontend. The toolkit also supports Speaker Verification, End-to-End Speech Translation, and Keyword Spotting. Recognized with the NAACL2022 Best Demo Award, PaddleSpeech aims to empower both industrial applications and academic research through its efficient, flexible, and scalable implementation, offering modules for training, inference, testing, and deployment.

Qwen-Audio (Qwen Large Audio Language Model) is an open-source multimodal AI tool from Alibaba Cloud, serving as a foundational audio-language model. It processes various audio types, including human speech, natural sounds, music, and songs, alongside text inputs, to generate text outputs. The tool is built on a multi-task learning framework, enabling knowledge sharing across over 30 tasks and supporting diverse audio-oriented scenarios. Qwen-Audio-Chat, an instruction fine-tuned version, offers multi-turn dialogues, flexible interaction with multiple audio inputs, and creative capabilities. It excels in benchmarks like Automatic Speech Recognition, Speech-to-text Translation, and Audio Question & Answer, making it a powerful tool for audio understanding and processing.

Catalyst.jl is a symbolic modeling package designed for the analysis and high-performance simulation of chemical reaction networks and related dynamical systems. It supports various simulation types including ODE, steady-state ODE, SDE, stochastic chemical kinetics (jump), and hybrid simulations. Models can be specified using an intuitive domain-specific language (DSL) or constructed programmatically. Built on ModelingToolkitBase.jl and Symbolics.jl, Catalyst leverages symbolic computation for tasks like sparsity exploitation, Jacobian construction, and dependency graph analysis. It integrates seamlessly with the broader Julia and SciML ecosystems for advanced analyses such as sensitivity analysis, parameter estimation, and bifurcation analysis, making it a powerful tool for researchers and developers in systems biology and scientific machine learning.