AlphaFold 3 is an AI model developed by Google's DeepMind team that can predict the three-dimensional structures of biomolecules such as proteins, nucleic acids (DNA and RNA), small molecules, ions, and modified residues. The model has achieved significant accuracy in structure prediction...
PDFMathTranslate is an open-source PDF document translation tool designed for translating PDF files such as scientific papers. It preserves the original layout, including formulas and charts. PDFMathTranslate supports bilingual comparison and maintains the original...
MoneyPrinterTurbo is an open-source AI short video generation tool that automatically generates video text, materials, subtitles, and background music based on user-provided video themes or keywords, creating high-definition short videos. The tool supports APIs...
iRAG is a text-based image enhancement technology launched by Baidu at the 2024 Baidu World Conference. It combines Baidu Search's massive image resources and powerful basic model capabilities to solve the challenges of large-scale modeling in text-based image processing...
Vidu 1.5 is the latest version of the AI video generation platform launched by Shengshu Technology, dedicated to helping creators express themselves freely and create efficiently. It features a multimodal video model, supports reference-generated videos, image-generated videos, and text-generated videos, ensuring...
AgentReview is a framework based on Large Language Models (LLM) that simulates the academic peer review process. Using LLM proxies, AgentReview simulates the roles of reviewers, authors, and area chairs, supporting researchers in respecting privacy...
CHANGER is an industrial-grade, supernatural AI head-swapping and chroma keying technology used in digital content creation to seamlessly integrate an actor's head onto a target body. It is suitable for visual effects, digital human creation, and virtual avatars. CHANGER is based on chroma keying technology...
Kiroku is a multi-agent system that assists users in organizing and writing documents. Inspired by its founder's academic writing experience during his doctoral studies at Stanford University, Kiroku simulates the interaction between a student and a mentor, helping users quickly generate...
Vision Search Assistant (VSA) is a framework that combines Visual Language Models (VLMs) and web agents to enhance the model's ability to understand unknown visual content. Based on internet retrieval, it enables VLMs to process and answer questions about unknown...
MVDrag3D is an innovative 3D editing framework that combines multi-view generation and reconstruction priors to achieve flexible and creative drag-and-drop editing. The framework uses a multi-view diffusion model as a generation prior to ensure consistency across multiple rendered views...
Chonkie is a lightweight, fast, and feature-rich RAG (Retrieval-Augmented Generation) chunking library designed for text processing. Chonkie supports various chunking methods based on tokens, words, sentences, and semantics...
MSQA (Multi-modal Situated Question Answering) is a large-scale multimodal situated reasoning dataset designed to improve the understanding and reasoning abilities of embodied AI agents in 3D scenes. The dataset contains 251K question-answer pairs, covering...
Excalidraw is an open-source online drawing tool with a clean, hand-drawn style and real-time collaboration features. Excalidraw runs entirely in the browser, requires no installation, supports multiple users editing the same drawing simultaneously, and provides end-to-end encryption...
RMBG-2.0 is BRIA AI's latest open-source image background removal model. Based on advanced AI technology, it achieves high-precision foreground and background separation, reaching state-of-the-art (SOTA) level. RMBG-2.0...
JanusFlow, a Janus series product from DeepSeek, is a model used for multimodal understanding and generation tasks. It integrates autoregressive language models with calibration flow techniques to achieve image understanding and generation within a single model. The framework is based on decoupling...
SWE-Kit is an open-source framework from Composio that simplifies the development process of AI agents in software engineering. SWE-Kit provides a headless IDE environment and AI-native tools for building custom-coded agents, supporting integration with various agent types...
Text Behind Image is an open-source online tool that allows users to add text behind people in images, creating visually impactful posters and social media images. Users can add custom text behind the main subject in an image...
Yun Jin Tian Zhang is a general-purpose large model developed by Caiyun Technology based on the DCFormer architecture. DCFormer achieves a 1.7-2 times performance improvement over the traditional Transformer model. Yun Jin Tian Zhang, based on a fictional world, imbues the novel with...
MikuDance is an animation generation technique based on a diffusion model, integrating hybrid motion dynamics to animate stylized character art. MikuDance utilizes hybrid motion modeling and hybrid controlled diffusion techniques to address high-dynamic motion and parameter...
MATRIX-Gen is a multi-agent simulation system developed by a research team from Shanghai Jiao Tong University and Oxford University. It simulates a society composed of over 1000 AI agents with independent identities and personalities, generating diverse and high-quality training data...
BodyTalk is an AI video dubbing tool launched by Panjaya. Based on generative AI technology, it converts the voices of people in videos into other languages, automatically adjusts the facial expressions and body movements of the people in the video, and naturally matches the new language...
Windsurf, developed by Codeium, is an AI programming tool featuring real-time collaboration, multi-system support, and powerful context-aware capabilities. Windsurf combines collaborative Copilots with independently operating Agents...
OmniParse is a data parsing platform that transforms unstructured data into structured, workable data, optimized for GenAI (large-scale language model) applications. OmniParse supports approximately 20 file types, including documents, tables, etc.
Promptim is an experimental AI suggestion optimization library that improves the suggestion performance of AI systems on specific tasks through automated processes. Users provide initial suggestions, datasets, and custom evaluators, and Promptim automatically runs optimization loops...