Toonflow is an open-source, one-stop AI short drama creation tool that automatically transforms novels and scripts into structured scripts, intelligent storyboards, character visuals, and animated videos through multi-agent collaboration, achieving the process of novel → script → storyboard → video...
JoyAI-Echo is an open-source long-form video generation framework launched by JD.com, designed specifically for minute-level multi-shot story generation. The framework utilizes a cross-modal memory library, memory-driven post-training, Director Agent conversational editing, and lightweight...
Ideogram 4 is Ideogram's first open-source text-to-image generation model, boasting 9.3 billion parameters and trained from scratch rather than fine-tuning existing models. The model is specifically designed for high-quality image generation, particularly in design, marketing, and more...
MAI-Transcribe-1.5 is a speech-to-text model developed by Microsoft's AI team. It supports 43 languages, features context-aware keyword biasing, and achieves the lowest word error rate in the FLEURS benchmark test...
MAI-Voice-2 is Microsoft's next-generation text-to-speech (TTS) model, and is Microsoft's most expressive and natural-sounding speech synthesis model to date. Compared to its predecessors, it excels in fidelity, language coverage, speaker consistency, and emotional range...
MAI-Code-1-Flash is a lightweight code generation model developed by Microsoft specifically for developer workflows, deeply integrated with GitHub Copilot. The model employs adaptive output length control technology, maintaining high accuracy while achieving maximum...
Odysseus is an open-source, self-hosted AI workspace that provides a chat interface similar to ChatGPT/Claude. The tools run completely locally, with data remaining on the local machine and zero privacy leaks.
PaddleOCR-VL-1.6 is a document parsing visual language model (VLM) launched by the Baidu PaddlePaddle team, and is the latest upgraded version of the PaddleOCR-VL series.
Bernini is an open-source unified video generation and editing framework from ByteDance, employing a two-stage decoupled architecture of MLLM semantic planning and DiT visual rendering. A multimodal large model understands instructions and plans semantic sketches, while Diffusion...
Science Skills is an open-source collection of scientific skills from Google DeepMind, designed to accelerate AI agent-driven scientific research workflows. The tools integrate AlphaGenome, AlphaFold Database, UniPro...
MAI-Thinking-1 is Microsoft's first self-developed advanced inference model. The model employs a sparse MoE architecture with 35B active parameters and approximately 1T total parameters, and was trained entirely from scratch on clean, commercially licensed data, without using any...
OpenClacky is an open-source AI agent developed by Li Yafei's team, supporting one-click installation and cross-platform operation. The tool significantly reduces usage costs through a streamlined toolset, intelligent context compression, and model routing, allowing professionals to run it around the clock...
Mellum2 is a new open-source machine learning model for software engineering systems developed by JetBrains. The model activates only 2.5B parameters per token, is optimized for AI workflows in software engineering, and supports code generation, intelligent routing, and more...
ECC (Everything Claude Code) is a cross-platform AI Agent performance optimization system. The system provides 63 professional agents and 249 on-demand agents for 7+ AI programming tools, including Claude Code, Codex, and Cursor...
Polar is an open-source agent reinforcement learning (ARL) training framework launched by NVIDIA. Its core innovation lies in the fact that it can be integrated with RL algorithms such as GRPO for training without modifying the internal code of existing agent frameworks.
Qwen3.7-Plus is a new generation of multimodal large-scale model launched by Tongyi Qianwen, unifying vision and language into an integrated intelligent agent foundation. The model can perceive real-world scenes, read screens and operate GUIs, generate models based on visual references, etc.
Hermes Desktop is a community edition desktop application for Hermes Agent, developed by sir1st. It packages the Python runtime, the hermes-agent core, and the hermes-web-ui visual interface into a single executable file...
Gamma-World is a multi-agent world model introduced by NVIDIA that can solve the problem of multiple agents interacting equally and globally in the same simulated world.
Dubbing v2 is an AI dubbing model developed by ElevenLabs that supports automatic translation and dubbing of 29 languages, preserving the original speaker's timbre and emotion. The model offers a dual workflow mode: Auto Dub for quick preview generation, and Dub...
Cosmos 3 is NVIDIA's first fully open-source, multimodal physics AI foundation model, which uses a hybrid Transformer architecture to integrate visual reasoning, world generation, and action prediction capabilities.
MiniMax M3 is a brand-new AI model from MiniMax, boasting leading programming, agent, and long text processing capabilities. The model employs an innovative MSA sparse attention architecture, supporting ultra-long contexts of up to 1MB, resulting in a 20x efficiency improvement.
Qwen-VLA is a general-purpose vision-language-action model developed by Tongyi Labs. It uses Qwen3.5-4B as the visual-language backbone and is paired with a DiT action decoder with 1.15B parameters. The model uses a unified action trajectory prediction framework to...
Webwright is an open-source, terminal-native web-based intelligent agent framework from Microsoft Research. With only about 1000 lines of code, it allows AI models to write Playwright code, execute bash commands, view logs, and iteratively correct errors within the terminal...
ControlFoley is an open-source, controllable video audio generation model from Xiaomi, addressing the controllability challenges in the V2A (Video-to-Action) domain. The model uniformly supports three types of video dubbing tasks: text-guided, text-controlled, and reference audio control, utilizing its self-developed spatiotemporal audio...