Research & Education

StableSR is an open-source image super-resolution tool that leverages diffusion prior to enhance real-world images. It is designed to improve image detail and clarity, offering capabilities like arbitrary upscaling, including examples beyond 4K resolution. The tool supports various configurations for training and testing, including DDIM with negative prompts and specific models for FaceSR. StableSR provides options for different color correction types and memory optimization for GPU usage. It has been integrated into popular platforms such as Hugging Face, OpenXLab, Replicate, ModelScope, and ComfyUI, making it accessible for researchers and developers. The project is actively maintained with updates, including support for SD-Turbo and improved performance.

SurfSense is an open-source, privacy-focused alternative to NotebookLM, designed for teams requiring no data limits and extensive customization. It allows users to add an unlimited amount of sources and notebooks, bypassing the restrictions found in other platforms. SurfSense supports over 100 LLMs via OpenAI spec & LiteLLM, and 6,000+ embedding models, ensuring no vendor lock-in. It integrates with 27+ external data sources like Google Drive, OneDrive, Dropbox, Notion, and Slack. The tool also features real-time multiplayer support for collaborative work, AI file sorting, and a desktop app with Quick Assist, General Assist, and Extreme Assist for AI assistance across applications. SurfSense can be self-hosted for full data control and privacy.

Sword Health delivers AI Care with clinical oversight, combining artificial intelligence with human clinical expertise to provide comprehensive health solutions. The platform offers specialized programs like Thrive for musculoskeletal pain, Bloom for women's health, Pulse for cardiometabolic health, and Mind for mental health. It also features Predict, an AI engine to identify and treat members at risk for avoidable surgeries, and Sword Intelligence for operational scaling. Trusted by employers and health plans, Sword aims to reduce healthcare costs and improve outcomes through personalized, accessible, and scalable care.

TextGrad is an open-source framework designed for automatic "differentiation" via text, leveraging large language models (LLMs) to backpropagate textual gradients. It provides a powerful and intuitive API, similar to PyTorch, allowing users to define loss functions and optimize them using text feedback. The tool supports optimizing various variables, including code snippets, solutions to problems, molecules, and prompts. Recent updates introduce a new engine based on LiteLLM, expanding compatibility to models like Bedrock, Together, and Gemini, and enabling caching. TextGrad has been published in Nature, highlighting its significance in the field.

DistillKit is an open-source toolkit designed for the knowledge distillation of large language models (LLMs). It offers flexible and production-ready solutions for both online and offline distillation workflows, including advanced logit compression techniques. The toolkit is battle-tested, powering the training of Arcee's popular open-source models like Virtuoso, SuperNova Medius, and Blitz. Key features include real-time teacher inference for online distillation, training from pre-captured teacher outputs with compression for offline distillation, and novel polynomial approximation, quantization, and bit-packing for logit compression. It also provides flexible loss functions, sparse and dense distribution support, and HuggingFace integration, making it a robust solution for LLM optimization.

Chappy is an AI-powered mobile application designed to assist students with their homework and quiz questions. By leveraging GPT-based artificial intelligence, Chappy allows users to simply take a photo of a problem, and the app will provide instant answers and detailed explanations. This ad-supported service is offered at no cost, making it an accessible tool for students seeking quick academic assistance. The app focuses on providing step-by-step solutions, aiming to help users understand the underlying concepts rather than just providing direct answers. Chappy Labs developed the app with a commitment to user privacy, outlining policies for the collection and use of personal information, such as Advertising ID, to ensure a better service experience.

Kagi Search is a privacy-focused search engine and suite of AI tools designed to reclaim the web from clutter and tracking. It offers ad-free search results tailored to user needs with powerful customization controls, anti-tracking, and filters. Beyond search, Kagi provides an AI Assistant that accesses multiple AI models while keeping user data private, an Orion Browser for zero-surveillance browsing, and Web Essentials like Kagi Translate, Kagi News, and Kagi Summarize. The platform is user-funded, ensuring that its incentives align with user needs rather than advertising. Kagi emphasizes a "human-first" design, integrating AI thoughtfully as an opt-in, on-demand, and user-controlled feature.

VibeThinker-1.5B is a 1.5B-parameter dense model developed by WeiboAI, challenging the conventional belief that small models inherently lack robust reasoning capabilities. It employs an innovative post-training methodology centered on the "Spectrum-to-Signal Principle (SSP)" to systematically enhance output diversity. This involves a "Two-Stage Diversity-Exploring Distillation" in the SFT phase and a "MaxEnt-Guided Policy Optimization (MGPO)" framework in the RL phase. VibeThinker-1.5B demonstrates superior reasoning in mathematical and coding tasks, outperforming larger models like DeepSeek R1 and achieving comparable results to GPT OSS-20B Medium, all at a significantly lower cost.

visualnav-transformer offers official code and checkpoint releases for mobile robot foundation models, including GNM, ViNT, and NoMaD. This open-source repository facilitates the training of these models with custom data and their deployment on robots like TurtleBot2/LoCoBot. It supports diverse, cross-embodiment training data and allows for efficient fine-tuning or adaptation to new robots and tasks. The tool provides scripts for processing datasets, training models, and deploying them for navigation and exploration, making it a comprehensive solution for robotics researchers and AI developers in the field of visual navigation.

AI Translator Now is an iOS mobile application designed to leverage cutting-edge AI for language translation, primarily catering to learning and academic needs. The tool focuses on enhancing user productivity by delivering accurate and current information identification through its translation capabilities. Its intelligent system is engineered to continuously learn and improve the quality of translations, making it a valuable asset for students and learners who require reliable language support in their studies. While the specific features beyond core translation are not detailed, its emphasis on academic use suggests a focus on precision and contextual understanding.

Synthetic Sciences offers an AI co-scientist platform designed to automate and accelerate scientific research workflows. It enables users to delegate tasks such as literature reviews, training models on cloud GPUs, designing experiments, and generating publication-ready LaTeX papers. The platform features four distinct modes—Core Research, SOTA Biology, Flywheel, and Write—each tailored for specific research objectives and execution styles. It provides a unified workspace for continuous research, integrating with various tools like GitHub, Hugging Face, and Weights & Biases, and offers persistent agent runtimes for long-running workflows with automatic checkpointing and elastic compute profiles.

Lufe AI Translator is an AI-powered bilingual web translation extension designed to simplify reading foreign language webpages, PDFs, and images. It leverages advanced AI engines like Gemini, OpenAI, and Claude to provide fast and accurate translations. Key features include bilingual reading for web content, auto-detection and AI translation of web content, personalized translation display styles, and support for multiple web translate engines. The tool also offers image translation with automatic language detection and text recognition, as well as PDF translation. It's particularly beneficial for language learners, students, and researchers who need to understand content in various languages efficiently.

Seamless.AI is the #1 rated AI sales software designed to empower businesses with accurate sales leads and automated outreach. It functions as a real-time search engine, leveraging AI to crawl the web and find, verify, and validate contact information for B2B professionals. The platform offers a Data Engine for prospecting, buyer intent, and CRM enrichment, an Engagement Hub for AI-powered emailing, calling, and social selling, and AI Agents for automating workflows like outbound, inbound, and customer success. Seamless.AI is trusted by over 1 million sales leaders and guarantees data accuracy with a credit-back protection policy.

Dcipher Insight Booster is an AI-powered platform designed to automate and accelerate research, analysis, and report generation. It leverages state-of-the-art agentic AI workflows and seamless orchestration to manage the entire insight lifecycle, from information gathering to decision-ready reporting. The tool significantly boosts efficiency by automating manual research, data wrangling, and synthesis, allowing users to process thousands of documents and data sources in multiple languages rapidly. Key features include scalable parallel processing with Research Agents, customizable research pipelines, interactive insight landscapes, trend radar, semantic search, and source-linked, auditable results. It supports rich data sources, custom knowledge bases, and offers AI-powered report generation with full source attribution, making it ideal for enterprise-scale, multilingual research.

q.e.d Science is an AI-powered platform designed to enhance scientific research, review, and decision-making through critical thinking. It allows users to break down academic papers into claims, expose underlying logic, and identify weaknesses or gaps grounded in the latest research. The tool helps researchers understand how their work scores against hundreds of other papers and highlights its unique contributions. Built by scientists for scientists, q.e.d Science aims to improve scientific integrity and provides constructive criticism for research evaluation. It ensures data privacy, with user research remaining confidential and not used for training models to reproduce or rephrase unpublished work.

The University at Buffalo is a leading institution in artificial intelligence and data science research and education. It emphasizes the application of AI for social good and provides a range of educational programs. The university hosts Empire AI, a significant statewide consortium and supercomputing center, fostering advanced research capabilities. The Institute for Artificial Intelligence and Data Science (IAD) at UB brings together researchers from diverse disciplines to drive advancements in AI, contributing to both theoretical understanding and practical applications. The university offers over 500 degree programs, including those in Artificial Intelligence, and is recognized as a top public university in New York for research activity.

Looppanel is an AI-powered UX research tool designed to accelerate qualitative data analysis and insight generation. It provides high-quality transcription, automatic note-taking, and AI analysis to transform raw research data into actionable insights. The platform centralizes all customer feedback into a searchable repository, allowing teams to quickly find answers and make evidence-based decisions. Looppanel also facilitates seamless sharing of video clips and reports across organizations, ensuring customer truth is heard. With features like live note-taking, clustering, and Google-like search, it helps eliminate guesswork and build on user insights 10x faster, without sacrificing quality or control. It also boasts enterprise-grade security with SOC2 Type II and GDPR compliance.

Paraspeech is an advanced AI audio tool designed for macOS, offering instant and private offline speech-to-text transcription. Optimized for Apple Silicon, it provides fast and accurate transcription in over 100 languages, ensuring privacy by keeping audio and text data on-device. The tool integrates seamlessly across all applications, from editors to chat apps, without requiring plugins. Key features include automatic punctuation and capitalization, and AI Rewriting to polish transcriptions, which can also be processed locally. Users can choose between monthly, yearly, or lifetime licenses, with a free trial available. It's built for efficiency, consuming minimal resources while running in the background.

GeoSocial Artificial Intelligence is a research group based at IT:U Interdisciplinary Transformation University Austria. Led by Prof. Bernd Resch, the group is dedicated to interdisciplinary research, with a strong emphasis on addressing real-world challenges through digital innovation. A key initiative is the "Climate Agents" project, which employs a co-creation approach to study the effects of climate change. This project involves students, teachers, and citizen scientists in participatory formats, where they become "citizen scientists" and create "StoryMaps" to present their findings and communicate science. This fosters an understanding of scientific processes and strengthens trust in science. The broader university offers master’s and PhD programs designed to tackle complex issues through digital innovation, making this research group an integral part of IT:U's mission.

Satlas, developed by AI2, is an innovative AI tool designed for exploring how our planet is changing through the analysis of AI-annotated satellite imagery. It provides a platform for users to gain insights into environmental changes and urban development by leveraging advanced artificial intelligence. This tool is particularly useful for researchers, policymakers, and anyone interested in monitoring global shifts, offering a unique perspective on our world's evolution. By processing vast amounts of satellite data, Satlas helps in understanding complex patterns and trends that might otherwise go unnoticed, making it a powerful resource for Earth observation and analysis.

Socratic Models Image Captioning is an AI tool available on Hugging Face that specializes in generating descriptive captions for images. It leverages Socratic Models in conjunction with BLOOM, a large language model, to produce relevant and informative captions. While the live website currently shows a runtime error, the tool's core functionality is designed for image captioning, making it valuable for AI research and educational purposes. It aims to provide a platform for exploring advanced AI capabilities in understanding and describing visual content.

Machine-Mindset is an open-source project developed through a collaboration between FarReel AI Lab and Peking University's Deep Research Institute, focusing on an MBTI exploration of Large Language Models. It offers 32 distinct MBTI models (16 Chinese, 16 English) built on Baichuan and LLaMA2 platforms, along with an extensive self-constructed MBTI dataset. The project aims to achieve personality alignment for various MBTI types in LLMs through multi-stage training, including pre-training, fine-tuning, and DPO training. Researchers can access all training datasets and models to foster innovation in the integration of large language models and psychology. The project also explores future directions such as MoE architecture, personalized LLM needs, and emotional companionship.

Goanie, also known as Acuity or Anie, is an AI-powered narrative integrity and media forensics engine designed to detect engineered narratives, emotional coercion, and 'zombie facts' in news articles and web content. It provides a 0-100 integrity score and a detailed forensic breakdown. The tool employs three core analysis vectors: Reality Anchoring, which cross-references claims against live search indexes; Tribal Engineering Detection, identifying in-group signaling and rhetorical traps; and Neuro-Linguistic Intent Analysis, which detects emotional manipulation and prescriptive language. Acuity is available as a Chrome extension, iOS app, Android app, and web scanner, making it accessible across multiple platforms for real-time analysis.

Baichuan2 is a new generation of open-source large language models developed by Baichuan Intelligent Technology. These models are trained on an extensive corpus of 2.6 trillion high-quality tokens, achieving state-of-the-art results across various authoritative Chinese, English, and multilingual benchmarks. The release includes 7B and 13B base and chat versions, with 4-bit quantization available for the chat models. Baichuan2 models are fully open for academic research and offer free commercial use upon application and official approval. The models demonstrate strong performance in general, legal, medical, mathematics, code, and multilingual translation domains, making them suitable for a wide range of NLP applications.