AI Agents & Automation

hum.ai is dedicated to building advanced multimodal foundation models designed for practical, real-world applications. Their core focus is on leveraging satellite remote sensing and ground truth data to train these models, aiming to develop Artificial General Intelligence (AGI) for a deeper understanding of the natural world. The technology developed by hum.ai is currently being utilized in critical sectors such as nature conservation, carbon dioxide removal initiatives, and by various government agencies. This positions hum.ai at the forefront of applying AI to solve complex environmental and scientific challenges, providing robust solutions for data analysis and predictive modeling in these domains.

EyeLock for Chrome Extension is a privacy-focused tool that enhances productivity and privacy by monitoring user attention via webcam. Designed for Chrome, it automatically mutes tabs or opens a designated 'safe' page when the user looks away from the screen, helping to prevent distractions and maintain privacy in various environments. All face and attention detection runs locally on the device using MediaPipe Face Landmarker, ensuring no cloud processing, external uploads, or account requirements. Users can customize look-away sensitivity, camera source, calibration, sounds, and behavior to suit their routine, making it adaptable as a distraction blocker, focus timer companion, or online meeting helper.

deep-learning-keras-tf-tutorial is an open-source project offering a comprehensive tutorial series for learning deep learning. It focuses on practical implementation using TensorFlow 2.0, Keras, and Python, making it suitable for beginners. The series covers a wide range of topics from fundamental concepts like activation functions and gradient descent to more advanced areas such as CNNs, transfer learning, word embeddings, and distributed training. Each topic is accompanied by code examples, allowing users to learn deep learning from scratch and build a solid foundation in the field.

ChatGPT History Manager is a Chrome plugin designed to significantly improve the way users manage their ChatGPT conversations. It provides essential features such as conversation grouping, allowing users to organize chats into folders by topic or project. The plugin also enables pinning of important conversations for quick access and offers a robust full-text search capability, including keyword positioning, to easily locate specific information within past interactions. This non-invasive tool automatically syncs with ChatGPT conversations, ensuring an organized and efficient way to retrieve and manage chat history, making it ideal for users who frequently engage with ChatGPT.

PYNQ-Classification is an open-source framework designed for the rapid deployment of embedded Convolutional Neural Network (CNN) applications on PYNQ platforms. It leverages Python on Zynq FPGA to accelerate CNN processing. The repository provides instructions for setting up Caffe and Theano dependencies, and includes demos for LeNet and CIFAR-10 models. Users can download a pre-configured SD card image or manually set up dependencies. The framework also guides on regenerating Vivado and Vivado HLS projects for implementing additional CNN models, making it a valuable resource for researchers and developers working with FPGA-based CNN acceleration.

Lolo is an innovative AI-powered food and calorie tracker designed to simplify dietary management. Users can log their meals by simply describing what they ate in plain text, bypassing traditional complicated drop-down lists and extensive food databases. The app leverages AI to accurately track food and manage calorie intake, making it easier to stay on top of dietary goals. Lolo is adaptable to various special diets, including those for fitness, diabetes, pregnancy, or other conditions, by adjusting its recommendations based on user-defined profiles. It also features AR scanning capabilities to quickly populate nutritional data from food labels, enhancing accuracy and user convenience. The app calculates daily caloric and nutrient needs using established models like the Mifflin-St Jeor Equation and guidelines from the U.S. Department of Health and Human Services.

Awesome-AGI-Agents is an open-source GitHub repository that provides a continuously updated, curated list of resources related to Artificial General Intelligence (AGI) agents. This comprehensive collection includes various types of content such as insightful articles and videos, academic papers, and cutting-edge projects like Auto-GPT and MetaGPT. It also features development platforms like LangChain and SuperAGI, making it a valuable hub for developers and researchers. The repository aims to consolidate key information and advancements in the AGI agent landscape, offering a centralized point for exploration and learning.

Kokoro-FastAPI is a robust, open-source text-to-speech solution built as a Dockerized FastAPI wrapper for the Kokoro-82M model. It supports multiple languages, including English, Japanese, and Chinese, with Vietnamese support planned. The tool offers both NVIDIA GPU accelerated PyTorch inference and CPU ONNX support, ensuring flexibility across different hardware setups. A key feature is its OpenAI-compatible Speech endpoint, simplifying integration into existing workflows. It also includes debug endpoints for system monitoring, an integrated web UI, and advanced capabilities like phoneme-based audio generation, per-word timestamped caption generation, and voice mixing with weighted combinations. The system automatically handles natural boundary detection for long-form text and provides streaming support for real-time audio output.

KeyboardGPT is an LSPosed Module designed to seamlessly integrate generative AI, such as ChatGPT, into your Android device's keyboard. This innovative tool allows users to access AI capabilities directly from any app, enhancing productivity and communication. It supports a wide range of keyboards, including Google Gboard and Microsoft Swiftkey, and is compatible with both rooted and unrooted devices running Android 16 or later. Key features include AI chat completions, text formatting (bold, italic, crossout, underline), and web search functionality, all accessible through simple commands within the keyboard interface. KeyboardGPT also supports various AI providers like Gemini, Groq, OpenRouter, and ChatGPT, offering flexibility in AI model choice.

ClawMetry is a free, open-source real-time observability dashboard specifically designed for AI agents built on OpenClaw. It offers comprehensive monitoring capabilities, including token cost tracking, cron job monitoring with failure alerts, and detailed visibility into sub-agents' activities like files, commands, tools, and thinking processes. Users can also track memory file changes and review session history with a timeline and cost breakdown. The dashboard provides cost breakdowns per-session, per-model, and per-tool, enabling users to identify and optimize expensive operations. ClawMetry runs locally, ensuring no cloud dependencies or telemetry, and supports both dark and light themes.

Macaw-LLM is an exploratory open-source project that pioneers multi-modal language modeling by seamlessly combining image, video, audio, and text data. Built upon the foundations of CLIP, Whisper, and LLaMA, it offers a unique approach to integrating diverse data types. Key features include simple and fast alignment to LLM embeddings, one-stage instruction fine-tuning, and a newly created multi-modal instruction dataset covering image and video modalities. The architecture leverages CLIP for image/video encoding, Whisper for audio encoding, and LLaMA (or Vicuna/Bloom) as the core language model. This tool is designed for researchers and developers to explore and advance the field of multi-modal AI.

Neuron Poker provides an open-source OpenAI Gym environment specifically designed for training neural networks to play Texas Hold'em poker. Leveraging Keras-RL for deep reinforcement learning, this tool offers features like virtual rendering to visualize gameplay and Monte Carlo simulations for accurate equity calculation. It supports various agent types, including random, keypress-controlled, equity-based, and Deep Q learning agents. The environment is highly customizable, allowing users to add their own player models and collaborate through pull requests. Advanced users can integrate a C++ version of the equity calculator for significantly faster computations, making it an ideal platform for AI researchers and developers focused on poker AI.

MCP Client for Ollama (ollmcp) is a powerful, interactive terminal application (TUI) designed for connecting local Ollama LLMs to one or more Model Context Protocol (MCP) servers. This client facilitates advanced tool use and workflow automation for developers. It offers a rich, user-friendly interface to manage tools, models, and server connections in real-time without requiring coding. Key features include agent mode for iterative tool execution, multi-server support, streaming responses, human-in-the-loop tool execution for safety, and advanced model configuration. It's built for developers working with local LLMs, streamlining their workflow with features like fuzzy autocomplete, hot-reloading for development, and comprehensive history management.

The NSF AI Institute for Artificial Intelligence and Fundamental Interactions (IAIFI) is a leading institute dedicated to pioneering interdisciplinary research at the intersection of AI and physics. It aims to advance fundamental physics knowledge, from the smallest building blocks of nature to the largest structures in the Universe, while simultaneously galvanizing AI research innovation. IAIFI focuses on developing AI approaches that incorporate first principles from physics and tackles challenging problems such as precision calculations and gravitational wave detection. Beyond research, IAIFI is committed to empowering the next generation of AI+Physics talent through various educational programs, including fellowships, summer schools, and workshops, and building a dynamic AI+Physics community through events and collaborations.

SentientAI is an AI product and services company dedicated to helping organizations become AI-first enterprises. The company focuses on transforming businesses by integrating advanced AI and data strategies into their core operations. SentientAI aims to empower businesses to leverage artificial intelligence for enhanced decision-making, operational efficiency, and strategic planning. By providing AI products and services, SentientAI assists organizations in becoming more "conscious" through the intelligent application of AI technologies.

R2R is an advanced, production-ready AI retrieval system designed for Agentic Retrieval-Augmented Generation (RAG). It provides a robust RESTful API for seamless integration into existing workflows. Key capabilities include multimodal content ingestion, allowing it to process various file types like .txt, .pdf, .json, .png, and .mp3. The system features hybrid search, combining semantic and keyword search with reciprocal rank fusion for highly relevant results. R2R also supports automatic entity and relationship extraction for knowledge graph creation, and includes a Deep Research API for multi-step reasoning to deliver context-aware answers. It's an open-source solution, making it accessible for developers to build sophisticated AI applications.

Stemgon is an IT consulting firm specializing in strategic IT consulting, cloud solutions, AI & Machine Learning, cybersecurity, digital transformation, and custom development. With over a decade of experience, Stemgon helps businesses align technology with their objectives, optimize cloud infrastructure, and implement intelligent automation. They offer comprehensive security solutions, end-to-end modernization of business processes, and bespoke software development to meet unique business requirements. Stemgon emphasizes a proven track record with over 500 successful projects, an expert team, 24/7 support, and scalable solutions, aiming for high client satisfaction.

Stock-Trading-Environment is an open-source project providing a custom OpenAI Gym environment designed for simulating stock trades using historical price data. This tool is ideal for developers, researchers, and quantitative analysts looking to build, test, and refine their AI-driven trading algorithms in a controlled and reproducible setting. By leveraging the OpenAI Gym framework, it offers a standardized interface for reinforcement learning agents to interact with a simulated market. The environment allows for backtesting strategies against real-world historical data, enabling users to evaluate performance and identify potential improvements before deployment in live markets. It's a valuable resource for anyone interested in applying machine learning to financial trading.

Pefai is an AI-powered platform designed to transform ideas into functional software solutions. It guides teams through the entire process, from initial ideation to technical definition, streamlining development. The platform specializes in auto-generating secure, no-code applications that are both traceable and scalable. Pefai aims to reinvent industries by providing a simple, quick, and affordable way to develop and deploy software, making advanced application creation accessible without extensive coding knowledge. This approach allows businesses to rapidly innovate and adapt to market demands.

PreciseRoIPooling is an open-source implementation of the Precise RoI Pooling (PrRoI Pooling) method, as proposed in the ECCV 2018 paper "Acquisition of Localization Confidence for Accurate Object Detection." This tool is designed to improve object detection accuracy by providing an integration-based average pooling method for RoI Pooling, which avoids quantization and offers a continuous gradient on bounding box coordinates. Unlike traditional RoI Pooling or RoI Align, PrRoI Pooling allows for the optimization of RoI coordinates through continuous gradients. The repository provides implementations for PyTorch (versions 1.0+ and 0.4) and TensorFlow (2.2), primarily supporting CUDA. It is a valuable resource for researchers and developers working on advanced object detection models.

Singa is an open-source distributed deep learning platform developed by Apache. It provides a flexible architecture for training deep learning models across various devices and distributed environments. The platform supports a wide range of deep learning models and offers tools for efficient computation and data management. Singa is particularly well-suited for researchers and developers who require a robust and scalable solution for their large-scale AI projects, enabling them to build, train, and deploy complex neural networks. Its open-source nature fosters community contributions and allows for extensive customization to meet specific project requirements.

WaiNSFWIllustrious V130, hosted on Hugging Face, provides a comprehensive platform for AI collaboration and compute. It offers a range of services including storage for models and datasets, hardware for running AI applications via Spaces, and Inference Endpoints for deploying ML models. Users can choose from PRO accounts for individuals, Team plans for growing teams, and Enterprise solutions for custom needs. The platform details various CPU and GPU options for Spaces and Inference Endpoints, with transparent hourly pricing. It also highlights features like private storage, inference credits, ZeroGPU access, and advanced organizational controls for team and enterprise users.

WaiNSFWIllustrious V130 Space is an AI image generation tool hosted on Hugging Face, enabling users to create images from textual descriptions. The platform provides various customization options, including the ability to specify a negative prompt to guide the AI away from unwanted elements, adjust image size and quality, and randomize the seed for diverse results. While the core functionality is accessible, advanced features and increased resource allocation are available through Hugging Face's PRO account and paid Spaces hardware options. This tool is marked as containing sensitive content, indicating its potential for generating NSFW or mature imagery.

Prompt Refine was a dedicated tool for enhancing prompt engineering workflows, allowing users to methodically improve their Large Language Model (LLM) prompts. It integrated with various AI models, including OpenAI, Anthropic, Together, and Cohere, providing a versatile environment for prompt development. Key functionalities included comprehensive history tracking to analyze and compare different prompt runs, enabling users to refine their approaches based on past results. The platform also supported the creation and reuse of variables within prompts, streamlining the experimentation process. Users could export their experiments to CSV for further analysis, making it a valuable asset for data-driven prompt optimization. However, the tool has since been shut down.