AI Agents & Automation

Aphra is an AI secretary designed to streamline digital interactions for freelancers, solopreneurs, and consultants. It proactively triages your inbox, drafts replies in your tone, and manages your calendar, all before you start your day. The tool also offers daily briefings, voice-to-email functionality for drafting replies on the go, and automatic detection and resolution of calendar conflicts. Aphra aims to reduce time spent on email management and scheduling, allowing users to focus on core work. It emphasizes data security with end-to-end encryption, SOC 2, and GDPR compliance, and supports major email providers like Gmail, Outlook, and Microsoft 365.

ThinkGPT is a Python library designed to augment Large Language Models (LLMs) by implementing Chain of Thoughts techniques. It enables LLMs to think, reason, and act as generative agents, addressing common limitations such as restricted context windows. Key features include memory management for LLMs to recall past experiences, self-refinement capabilities to improve model-generated content, and knowledge compression techniques to fit extensive information within an LLM's context. The library also offers inference based on available data, natural language conditions for decision-making, and efficient context length management, all through an easy-to-use Pythonic API.

AI-FORWARD is a Paris-based AI consulting firm specializing in assisting Small and Medium-sized Enterprises (SMEs) and Intermediate-sized Enterprises (ETIs) with their artificial intelligence initiatives. The firm provides a comprehensive suite of services, including strategic diagnostics to identify AI opportunities, practical deployment of AI solutions, and certified Qualiopi training programs. AI-FORWARD also emphasizes responsible AI governance, ensuring ethical and effective integration of AI technologies. With a track record of accompanying over 130 companies and achieving 99% client satisfaction, AI-FORWARD aims to transform AI into measurable business results for its clients.

Dwarf Land offers an immersive, real-time simulation of an autonomous dwarf civilization, powered by a sophisticated multi-tier AI system. Users can observe 29 unique cities across a vast, procedurally generated world, each inhabited by citizens making decisions based on their needs, social interactions, and strategic planning. This platform allows for a unique interactive experience where individuals can sponsor dwarves to upgrade their AI reasoning, influencing their intelligence and decision-making capabilities within the evolving ecosystem. It's designed for enthusiasts of AI simulations, virtual worlds, and strategic gaming who wish to explore emergent behaviors and complex AI interactions in a dynamic, living environment. Dwarf Land provides a fascinating sandbox for observing how AI agents can create and manage a thriving, autonomous society, with features like crafting, trading, travel, and infrastructure development.

Dragonfruit AI is an all-in-one enterprise AI platform specifically designed for retail, leveraging existing camera infrastructure to provide actionable intelligence. It employs computer vision and specialized AI agents to address critical retail functions such as shoplifting detection, queue management, checkout loss prevention, and customer journey insights. The platform offers a unified dashboard for centralized control across various applications and agents, making it easy for LP, Operations, and CX teams to manage. Dragonfruit AI is built for scalability and cost-effectiveness, integrating with existing VMS and camera systems even in low-bandwidth environments. Its patented split AI architecture focuses on edge-first processing to reduce bandwidth and cloud compute costs, making it an efficient solution for multi-location enterprises.

The Russian LLM Leaderboard is a platform hosted on Hugging Face designed for the evaluation and comparison of Russian language models. It enables users to submit their language models for assessment and monitor their performance relative to other models on the leaderboard. The platform provides a structured environment for benchmarking AI task automation and chatbot capabilities specifically within the Russian language context. By offering a centralized space for model evaluation, it helps developers and researchers understand the strengths and weaknesses of various Russian LLMs, fostering competition and improvement in the field. The tool is open source, promoting transparency and community contribution to the evaluation process.

Three Sigma is an AI research tool designed to streamline document interaction and utilization. It significantly reduces reading time, claiming up to a 90% reduction, by answering questions directly from your documents. The platform supports various document formats, making it versatile for different types of content. Future integrations include GPT-4 for enhanced image understanding, further expanding its capabilities. This tool aims to simplify the process of extracting information and insights from extensive documentation, making it an efficient solution for research and information retrieval.

Russian Text To Speech is a web-based AI tool developed by TeraTTS, available on Hugging Face, designed to convert Russian text into spoken audio. Users can input any Russian text and choose from various voice models to generate speech. A key feature is the ability to optionally add correct stress marks and the letter 'ё' to the text, enhancing the accuracy and naturalness of the generated audio. Furthermore, the application allows users to adjust the length scale, making the speech sound longer or shorter as needed. This tool is ideal for creating educational materials, developing voice applications, or generating narrations in Russian.

awesome-mixture-of-experts is a comprehensive GitHub repository dedicated to curating resources on Mixture-of-Experts (MoE) models in deep learning. It serves as a valuable collection of papers, code, and other relevant materials for anyone interested in this advanced AI architecture. The repository is organized into sections covering open models, must-read papers, MoE model publications, MoE system publications, MoE application publications, and libraries. It features prominent MoE models like DeepSeekMoE, LLaMA-MoE, and Mixtral of Experts, alongside foundational and recent research papers. This resource is ideal for researchers, data scientists, and developers looking to explore, understand, and implement MoE models.

Step-Audio 2 is an end-to-end multi-modal large language model developed by stepfun-ai, focusing on industry-strength audio understanding and speech conversation. It excels in advanced speech and audio understanding, comprehending and reasoning semantic, paralinguistic, and non-vocal information. The model facilitates intelligent speech conversations that are contextually appropriate and can analyze user paralinguistic information like age and emotion for more accurate interpretations. Step-Audio 2 also supports tool calling and multimodal RAG, allowing it to access real-world knowledge and generate responses with fewer hallucinations, even switching timbres based on retrieved speech. It demonstrates state-of-the-art performance across various audio understanding and conversational benchmarks, with mini versions available under an Apache 2.0 license.

Claude35Sonnet🤖 is an AI chatbot built on Hugging Face Spaces, offering an interactive platform to engage with the Claude AI model. Users can chat with Claude, receive responses, and conveniently save their conversations as files for future reference. Beyond just chatting, the tool includes robust file management features, allowing users to organize their saved interactions. Additionally, it provides galleries for viewing images, videos, and audio, enhancing the overall user experience by integrating multimedia capabilities directly within the chat environment. This makes it a versatile tool for both conversational AI and basic media management.

Scaling FineWeb is an AI research tool designed to evaluate multilingual models across a vast array of over 1000 languages. This tool, hosted on Hugging Face, utilizes a comprehensive suite of evaluation tasks known as FineTasks to assess model performance. It is particularly useful for researchers and developers working on multilingual AI development and natural language processing (NLP) research. By providing a structured approach to finding signals in hundreds of evaluation tasks, Scaling FineWeb enables users to gain insights into how models perform in diverse linguistic contexts, facilitating the improvement and scaling of AI technologies globally.

Frigate is a comprehensive, local NVR solution specifically designed for integration with Home Assistant, featuring advanced AI object detection capabilities. It leverages OpenCV and TensorFlow to perform real-time object detection directly on local IP camera feeds. The system is engineered for minimal resource consumption and maximum performance, employing low-overhead motion detection to trigger object detection only when necessary. Frigate utilizes multiprocessing to ensure real-time processing and communicates via MQTT for seamless integration with other systems. It supports 24/7 recording with retention settings based on detected objects, re-streaming via RTSP, and offers WebRTC & MSE for low-latency live viewing. Use of a GPU or AI accelerator is strongly advised for optimal performance.

MEGVII旷视 is a leading Chinese AI company specializing in full-stack AIoT solutions. The company integrates advanced algorithms, software, and hardware to create comprehensive systems for various applications. Its core offering includes the AI productivity platform Brain++, which comprises MegEngine for algorithm training and deployment, MegCompute for shared and distributed computing power, and MegData for data processing and management. MEGVII旷视 focuses on three main scenarios: consumer IoT, city IoT, and supply chain IoT, providing validated industry solutions to enhance efficiency and user experience. Their product range includes AIoT application computing integrated machines, intelligent servers, analysis boxes, facial recognition access control systems, and smart network cameras, all designed to make the physical world smarter and more connected.

Shyguy's Wingman is an interactive AI chatbot game designed to assist a shy character in navigating social interactions and securing a date. Players take on the role of a 'Wingman' helper, engaging in conversations with various characters through either text or voice input. The core gameplay involves gathering information, making strategic decisions, and guiding the shy protagonist, Shyguy, to successfully talk to Jessica. Built for the Mistral AI Game Jam, this tool offers an engaging experience in AI-driven conversations and interactive storytelling, allowing users to influence the narrative through their choices.

Zephyr Playground is an AI chatbot platform designed for experimenting with and testing various AI models. It provides a space for users to interact with and explore the capabilities of AI, making it suitable for educational purposes and for developers looking to test different AI functionalities. The platform allows for hands-on experience with AI technology, offering a practical environment for learning and development. While the live website indicates a runtime error and scheduling failure due to hardware capacity, the intention of the tool is to provide a playground for AI model interaction.

Yehia 7B Preview is an AI chatbot available on Hugging Face Spaces, designed to provide thoughtful and fluent responses to questions or topics entered in Arabic. This tool allows users to interact with the Yehia 7B language model, offering a platform for testing and evaluating its capabilities. Users have control over various output parameters, including the length of the reply, the level of creativity in the response, and the repetition settings, enabling a customized conversational experience. It is suitable for individuals interested in exploring AI language models, particularly those focused on Arabic natural language processing, for research, development, or educational purposes.

TileRT is an open-source, tile-based runtime engineered for ultra-low-latency Large Language Model (LLM) inference. It aims to push the boundaries of LLM latency without compromising model size or quality, allowing models with hundreds of billions of parameters to achieve millisecond-level time per output token (TPOT). Unlike traditional inference systems optimized for high-throughput batch processing, TileRT prioritizes responsiveness, making it ideal for applications like high-frequency trading, interactive AI, real-time decision-making, and AI-assisted coding. It achieves this by decomposing LLM operators into fine-grained tile-level tasks and dynamically rescheduling computation, I/O, and communication across multiple devices to minimize idle time and improve hardware utilization. TileRT currently supports models like GLM-5 and DeepSeek-V3.2 and offers Multi-Token Prediction (MTP) for efficient longer output generation.

AND Solutions Pte. Ltd. offers a comprehensive suite of AI-powered software solutions designed for banks and financial institutions. Their platform integrates lending, document intelligence, and credit decisioning into a single connected system. Key products include Looms for end-to-end loan origination and management with automated workflows and flexible loan configurations, Mindox for intelligent document processing to reduce manual tasks and enhance data accuracy, and advanced credit scoring tools like Scorecard Builder and Custom AI Scoring for accurate risk assessment and automated approvals. The solutions are built to streamline operations, accelerate lending, and improve decision-making for financial institutions across Southeast Asia.

Table Transformer (TATR) is a deep learning model developed by Microsoft for extracting tables from unstructured documents, including PDFs and images. Based on object detection, TATR can be trained to work across various document domains, with pre-trained model weights available for the PubTables-1M dataset. The repository also provides the official code for the PubTables-1M dataset, a large-scale dataset for table detection, structure recognition, and functional analysis, and the GriTS evaluation metric for table structure recognition. Researchers and developers can use TATR to detect and recognize tables, convert them to HTML or CSV, and train custom models for specific needs.

XingTong is a user-friendly AI tool available as a Hugging Face Space, designed for voice modification of audio files. Users can upload a clean audio file, provided it is less than 90 seconds in duration, and then choose from a selection of desired voices. The application processes the input and returns the audio with the altered voice. This tool is ideal for individuals looking to experiment with voice changes for creative projects, content creation, or personal use without requiring complex audio editing software. Its straightforward interface makes it accessible for non-technical users.

2V is a platform designed for creating interactive AI experiences, focusing on personalization. Users can build AI interactions that reflect their own life and personality, leading to unique and engaging AI simulations. The platform aims to provide a distinct approach to AI interaction, moving beyond generic responses to offer a more tailored and immersive experience. While specific features are not detailed, the core offering revolves around enabling users to craft AI agents that embody personal characteristics and respond in a customized manner, fostering a deeper connection with the AI.

Trankit is a light-weight, transformer-based Python toolkit designed for multilingual Natural Language Processing (NLP). It offers a trainable pipeline for fundamental NLP tasks across more than 100 languages, and includes 90 downloadable pretrained pipelines for 56 languages. Trankit outperforms other state-of-the-art multilingual toolkits like Stanza in various tasks, including sentence segmentation and dependency parsing, while maintaining efficiency in memory usage and speed. Key features include an Auto Mode for automatic language detection, a command-line interface for ease of use, and support for tasks such as tokenization, part-of-speech tagging, morphological feature tagging, dependency parsing, and named entity recognition. It also allows users to build and share customized pipelines.

Zephyr Chat is an AI chatbot hosted on Hugging Face Spaces, designed for conversational interactions. Users can engage with an AI model to get answers to questions or simply start conversations. The platform offers various examples to help users get started and explore the AI's capabilities. While the Space is currently paused, it aims to provide an accessible way to experience AI chat technology. It's a straightforward tool for anyone interested in interacting with an AI model in a chat format.