AI Agents & Automation

ScaleCUA is an open-source project that introduces a new approach to scaling computer use agents (CUAs) with cross-platform data. It offers a large-scale dataset spanning six operating systems (Windows, macOS, Ubuntu, Android) and three GUI-centric task domains, collected via a closed-loop pipeline combining automated agents and human experts. The project includes ScaleCUA-Models, a general-purpose agent capable of GUI-centric task completion across various environments, and a comprehensive SFT Codebase for training agents based on Qwen2.5-VL and InternVL. Additionally, ScaleCUA provides an interactive playground with realistic environments for Ubuntu, Android, and Web, and an online evaluation suite to benchmark agent capabilities.

Scene-Graph-Benchmark.pytorch is a comprehensive open-source codebase for Scene Graph Generation (SGG) methods, built on top of the well-known maskrcnn-benchmark project. It serves as a PyTorch implementation of the paper "Unbiased Scene Graph Generation from Biased Training CVPR 2020." The tool allows users to visualize and extract scene graphs from custom images and datasets, supporting various SGG protocols including Predicate Classification (PredCls), Scene Graph Classification (SGCls), and Scene Graph Detection (SGDet). It integrates predefined models like Neural-MOTIFS, IMP, VCTree, and Transformer, and offers options for customizing models. The project also introduces and clarifies SGG metrics, achieving state-of-the-art Recall@k on SGCls & SGGen. It's designed to be novice-friendly and easier to modify than previous frameworks.

paz is a hierarchical perception library built in Python, designed for autonomous systems. It offers a comprehensive suite of functionalities for computer vision tasks, including pose estimation, object detection, instance segmentation, keypoint estimation, and face recognition. The library is built on Tensorflow 2.0, OpenCV, and NumPy, providing a robust framework for developers. paz features a hierarchical API structure with high-level functions for out-of-the-box predictions, mid-level APIs for building custom pipelines, and low-level backend functions for fine-grained control. It also includes built-in messages for data exchange with other frameworks like ROS, custom callbacks for training evaluation, and data loaders for multiple datasets such as OpenImages and VOC. The library implements various models that can be retrained with custom data, making it a versatile tool for researchers and developers in robotics and AI.

GPU Poor LLM Arena is a platform designed for the comparison and evaluation of compact language models, specifically those with up to 14 billion parameters. It offers a battle arena format where users can input a text prompt and receive side-by-side answers from two different language models. This setup facilitates direct comparison, allowing users to vote for the better reply and contribute to a community-driven ranking. The tool is ideal for researchers, developers, and enthusiasts interested in understanding the practical performance of smaller, more resource-efficient AI models without requiring extensive GPU resources. It provides insights into the capabilities of frugal AI options.

SynRiva positions itself as an AI solution provider dedicated to creating a smarter future through next-generation AI solutions. While the specific offerings are not yet detailed, the company's website currently serves as a placeholder, stating that "Something Great Is On The Way." This suggests an upcoming launch of innovative AI products or services. The site includes a contact form for inquiries and an option to sign up for email updates, indicating an intent to engage with potential users and keep them informed about their developments. The company is copyrighted to 2025, hinting at a future release or significant update.

SRSWTI is a Knowledge and Inference Platform designed to facilitate the understanding and application of knowledge. While specific features are not detailed on the publicly available pages, the platform's core offering revolves around providing tools and resources for knowledge management and inference. It aims to support users in various contexts, likely including educational and research environments, by enabling them to process, organize, and derive insights from information. The platform's focus on "Knowledge and Inference" suggests capabilities related to data analysis, pattern recognition, and potentially predictive modeling, catering to those who need to manage and leverage complex information effectively.

Limit is a digital insurance platform designed to streamline the process of obtaining cyber, professional, and management liability insurance. It offers instant quotes, recommendations, and comparisons from over 50 markets, eliminating the need for tedious follow-ups and email chains. The platform provides two main services: Bridge, which offers direct connectivity for instantly bindable quotes with 0% commission for agents using their appointed markets, and Wholesale, which gives access to admitted and E&S options through a team of expert wholesale brokers. Limit also features a Cyber Price Index, providing real-time pricing data to bring transparency to the insurance market. The service is free to sign up, with no subscription fees, minimum contracts, or premium requirements, making it accessible for agents to optimize their sales and client offerings.

YAKE! (Yet Another Keyword Extractor) is a lightweight, unsupervised automatic keyword extraction method designed to identify the most important keywords from a single document. It leverages text statistical features and requires no prior training, external corpus, or dictionaries, making it highly adaptable across various languages and domains, regardless of text size. Key features include its unsupervised approach, language and domain independence, and a focus on single-document processing. YAKE also offers keyword lemmatization to aggregate morphological variations and a text highlighting feature to mark extracted keywords within the original text. It can be used via command line or Python, offering flexibility for developers and researchers.

WriteMage is an AI application designed to seamlessly integrate ChatGPT capabilities into macOS and iOS devices. It allows users to interact with AI directly within any macOS app, functioning like Apple Spotlight with context-awareness, eliminating the need for copy-pasting. The tool features memory, enabling it to remember conversation context within a session, and saves all chat history locally with labels for easy retrieval. WriteMage also includes a Prompt Editor GUI, empowering users to customize and create their own prompts. For iOS, it integrates natively with the keyboard, working across various apps. The app offers both subscription and lifetime deal options, with a beta period that includes free usage and discounts.

OpenInfer is a full-stack enterprise inference infrastructure designed to run AI models anywhere, from edge devices to on-premise servers or the cloud, without hardware compromises. It connects distributed and heterogeneous compute resources, including CPUs, GPUs, and NPUs, into a single coordinated AI system. This approach allows AI to run where data lives, maximizing ROI, ensuring data sovereignty, and providing reliable, always-on performance. OpenInfer is built for simple deployment and maintainability, supporting collaborative AI by keeping agents, operations, and systems in sync for mission-critical and sovereign environments. It unlocks inference on fragmented compute, enabling large model inference where it previously couldn't operate, and is proven to deliver significant performance improvements on commodity hardware.

PandaGPT is a pioneering open-source AI model that excels in instruction-following across six distinct modalities, including visual and auditory inputs. It is the first foundation model to achieve this without explicit supervision, demonstrating advanced capabilities in complex understanding, reasoning, knowledge-grounded descriptions, and multi-turn conversations. Researchers and developers can leverage PandaGPT to perform intricate tasks such as detailed image description generation, creating stories inspired by videos, and answering questions based on audio. Its unique ability to process and semantically compose multimodal inputs, like connecting visual and auditory information, makes it a powerful tool for advanced AI research and application development. The project provides resources for running demos, training custom models, and accessing pre-trained checkpoints.

VizGPT is an AI-powered data visualization tool designed to help users create contextual data visualizations from tabular datasets through a conversational chat interface. It allows users to generate visualizations using natural language queries and then refine them step-by-step within the chat context, eliminating the need to retype complex queries. This approach makes data visualization accessible even for those unfamiliar with traditional tools like Tableau or pygwalker, which often require specific configurations and data transformations. VizGPT focuses on text-based visual exploration, enabling users to discover insights and ask new questions based on their findings. Key features include natural language to data visualization, chat-based editing, step-by-step data exploration, and the ability to upload custom CSV datasets. While strong in visualization, it's noted that data transformation and preparation are better handled by other tools.

The vonage-php-sdk-core is a robust PHP client library designed to facilitate seamless integration with the Vonage API. It provides comprehensive support for a wide range of communication services, including SMS, Voice, and Text-to-Speech. Developers can leverage this library to implement features such as number verification (2FA), sending messages across various platforms like WhatsApp, MMS, and Viber, and managing inbound messages via webhooks. The library requires a minimum PHP version of 8.1 and is easily installable via Composer. It offers flexible authentication options, including basic API key/secret and signature-based credentials, and allows for custom API endpoint configurations. The SDK also includes functionalities for verifying incoming message signatures, ensuring secure communication within applications.

VideoLLaMA3 is an open-source project offering a series of multimodal foundation models designed for advanced image and video understanding. It provides models like VideoLLaMA3-7B and VideoLLaMA3-2B, which are capable of tasks ranging from general image and video comprehension to more specialized applications such as multi-image comparison, visual referring, and grounding. The project includes detailed instructions for inference, training, and evaluation, making it suitable for researchers and developers. It supports various benchmarks for performance assessment and offers a flexible framework for preparing custom training data. The models are available on Hugging Face, facilitating easy access and integration into AI development workflows.

mpathic is an AI platform dedicated to human-centered AI safety, offering end-to-end evaluation across the model lifecycle. It provides tools for policy development, curation of expert ground truth datasets, and real-time feedback loops. The platform helps AI builders train, evaluate, and calibrate models for nuanced human behavior and high-risk environments, ensuring safe performance in real-world applications. Key features include human data creation by mpathic Experts for red teaming and benchmarking, Observing Agents for comprehensive conversational data analysis, and mpathic Studio for recording, analyzing, and improving high-risk interactions with structured feedback.

Narrative AI empowers sales leaders and their teams with AI-driven insights, comprehensive research, and personalized messaging tools to significantly improve account targeting and outreach efficiency. It helps sales professionals focus on high-value accounts by automating the creation of org charts, researching prospects, and drafting tailored messages and proposals. The platform integrates with existing CRMs like HubSpot and LinkedIn, layering intelligence on top of current workflows without requiring data migration. By reducing the time spent on manual research and drafting, Narrative AI allows sales teams to dedicate more effort to actual client conversations and strategic planning, ultimately boosting sales productivity and performance.

UI-TARS-desktop is an open-source multimodal AI agent stack designed to connect various AI models and agent infrastructure, enabling the creation of sophisticated GUI agents. This tool is particularly useful for integrating vision capabilities across different platforms, allowing for the development of AI-driven automated workflows. It provides a robust framework for developers to build and deploy intelligent applications, leveraging advanced AI functionalities to automate complex tasks and enhance user interfaces. The platform supports a wide range of features for managing code changes, automating workflows, and securing applications, making it a comprehensive solution for modern software development.

cursor-deepseek is a high-performance HTTP/2-enabled proxy server designed to bridge the gap between Cursor IDE's Composer and various powerful language models like DeepSeek, OpenRouter, and Ollama. This proxy translates OpenAI-compatible API requests into the specific API formats required by DeepSeek, OpenRouter, or Ollama, allowing Cursor's Composer and other OpenAI API-compatible tools to seamlessly integrate and utilize these models. Key features include HTTP/2 support for improved performance, full CORS support, streaming responses, function calling/tools, automatic message format conversion, and compression support. It also offers API key validation for secure access and Docker container support for easy deployment.

siml is an open-source repository offering popular Machine Learning algorithms implemented from scratch. It is primarily intended for educational use, accompanying blog posts that delve into the mathematical foundations and interpretation of these algorithms. The project aims to simplify complex academic literature, presenting ML concepts with straightforward mathematics and concise code. It includes notebooks explaining various algorithms such as Linear and Logistic Regression, Naive Bayes, Perceptron Classification, and applications of Wavelet Transform for signal analysis and classification. Users can install siml via pip or by cloning the repository, making it accessible for those looking to learn and experiment with fundamental ML concepts.

cosine_metric_learning offers a repository with code for training a metric feature representation, specifically tailored for person re-identification tasks. This tool is intended to be used in conjunction with the deep_sort tracker, implementing the approach described in the 'Deep Cosine Metric Learning for Person Re-identification' paper. It includes functionalities to train models on datasets like Market1501 and MARS, with options for different loss modes such as cosine-softmax. Users can monitor training progress and evaluation metrics using TensorBoard, export features for testing, and freeze trained models for deployment with Deep SORT. The repository provides detailed instructions for setting up datasets, initiating training, and evaluating model performance.

TripMate is an AI-driven travel planning tool designed to simplify the process of creating personalized and detailed itineraries. Leveraging advanced AI algorithms, it provides users with tailored travel plans, taking into account individual preferences and requirements. Key features include real-time recommendations for attractions, dining, and activities, ensuring an up-to-date and relevant travel experience. The tool also incorporates dynamic pricing, which can help users find cost-effective options for flights and accommodations. TripMate aims to streamline trip organization, making it easier for travelers to manage their plans and discover new destinations.

Trener Robotics develops Acteris, an AI software platform designed to transform standard industrial robots into intelligent, self-learning operators. Powered by Physical AI, Acteris enables robots to see, learn, and adapt to complex, high-mix, low-volume production scenarios where traditional automation often fails. The platform reduces deployment times, simplifies changeovers without reprogramming, and enhances robot autonomy with built-in recovery flows. It offers operational visibility through real-time production monitoring and is compatible with leading robot brands like ABB, Universal Robots, and FANUC, with plans for further expansion. Acteris addresses the challenges of rigid programming and high changeover costs in modern manufacturing.

Sftwtrs.ai is a cutting-edge AI lab specializing in delivering innovative solutions across various domains including security, predictive maintenance, demand forecasting, and personalized experiences. They aim to transform perceptions of AI by harnessing its potential to streamline business operations and drive growth. The company provides custom-built AI solutions, helping businesses scale and reduce costs through smarter automation. They emphasize cost-effective, high-quality solutions and offer continuous AI support to ensure solutions evolve with business needs. Sftwtrs.ai positions itself as an approachable lab for businesses seeking advanced AI capabilities without the overhead of building an in-house team.

FrontLogix offers advanced workforce management services and customer experience outsourcing, blending BPO expertise with WFM innovation. The platform utilizes AI to streamline customer engagement, improve response times, and personalize interactions, leading to stronger customer relationships and increased retention rates. Key offerings include WFM-managed services for staffing and scheduling, AI-enabled customer service, and comprehensive training and development programs for call center agents. FrontLogix also provides quality management solutions to ensure high standards in customer care and offers digital customer support across various channels. Their approach focuses on tailored solutions, leveraging customer insights data to identify patterns and optimize the customer journey.