Open-source audit reveals Claude Code's hidden timezone tracking, targeting 'Asia/Shanghai' and 'Asia/Urumqi', likely to restrict features or content for certain users, raising concerns over covert compliance censorship in AI tools.....
The U.S. Department of Commerce lifted export controls on the Anthropic Claude Fable5 and Mythos5 models on June 30th, allowing full access starting July 1st. Previously, due to a vulnerability in Fable5 that could be exploited to provide information about network attacks, the department had temporarily restricted access to the model on June 12th, causing a three-week downtime. The ban also affected Anthropic's foreign employees.
US Commerce Department lifts export controls on Anthropic's 'Fable 5' and 'Mythos 5' AI models, allowing license-free export, re-export, and domestic technology transfer. This marks a sharp policy shift just weeks after the White House forced suspension of overseas operations over technology leakage concerns.....
On June 30, Cainiao launched a customized full-chain logistics solution for tech going global, offering end-to-end services to AI computing, smart manufacturing, robotics and other tech firms. With surging demand for cutting-edge hardware and computing clusters, tech exports are expanding from consumer electronics to high-end productive tools. Cainiao thus focuses on high-end equipment sectors to support globalization.....
Create production-ready animations directly in Figma, export code or GIFs without additional tools.
StarVeil AI can convert stories into high-definition web comics. The AI automatically generates storyboards and supports multi-style 4K export.
Free tool, supports comment export from multiple platforms, including AI analysis and lottery functions
Free online AI tool to remove models or hangers, reconstruct the neckline, and export clothing pictures with a white background.
ryandono
This model is a quantized (Int8 / Q8) version of mixedbread-ai/mxbai-edge-colbert-v0-17m, exported specifically for use with transformers.js. It includes a custom projection layer required for the 48-dimensional ColBERT architecture, aiming to provide efficient vector retrieval capabilities.
taobao-mnn
Qwen3-VL-8B-Thinking-MNN is a 4-bit quantized version of the MNN model exported from Qwen3-VL-8B-Thinking, optimized for text generation and chat scenarios, and supports visual language understanding tasks.
A 4-bit quantized version of the MNN model exported from the gpt-oss-20b model, optimized for text generation tasks, supporting application scenarios such as chatting, with efficient inference performance and low memory usage
An MNN model exported from Qwen3-30B-A3B, featuring 4-bit quantization for efficient inference.
A 4-bit quantized version of the MNN model exported from Qwen2.5-Omni-3B, implemented through the llmexport tool, providing efficient inference performance.
A 4-bit quantized version of the MNN model exported from Qwen2.5-Omni-7B, optimized for text generation and chat scenarios, with efficient inference performance.
gokaygokay
FLUX.1-dev is a text-to-image model based on double exposure technology, specializing in the artistic fusion of human silhouettes with various scenes.
erax-ai
Efficient NSFW content detection model, suitable for image/video moderation or restricting minors' exposure to harmful content
glif-loradex-trainer
A LoRA model fine-tuned based on the FLUX.1-dev foundation model, specializing in generating dynamic scene images with long exposure effects
Qwen2-0.5B-Instruct-MNN is a 4-bit quantized version of the MNN model exported from Qwen2-0.5B-Instruct, suitable for text generation and chat scenarios.
A 4-bit quantized version of the MNN model exported from Qwen2.5-0.5B-Instruct using the llmexport tool, optimized for text generation and chat scenarios, featuring lightweight design and efficient inference.
ntc-ai
This is a LoRA model based on Stable Diffusion XL, specifically designed for generating images with long exposure photography styles.
deepghs
The officially leaked model from NovalAI, which can be used for general model training or AI drawing, has been exported to a format supported by Diffusers.
echarlaix
This model is a Vision Transformer (ViT) fine-tuned on the Food-101 dataset, statically quantized to INT8 using the Optimum tool and exported in OpenVINO Intermediate Representation format, suitable for efficient image classification tasks.
ONNX export version of T5-small model, supporting multilingual text generation and translation tasks.
FastAPI-MCP is a library that exposes FastAPI endpoints as Model Context Protocol (MCP) tools, with built-in authentication, supporting flexible deployment and ASGI transmission.
A feature-rich Markdown typesetting tool that supports multiple style themes and platform adaptation, providing real-time editing preview, image export, and API integration capabilities
Blender MCP VXAI is a powerful integration tool that allows users to control Blender through natural language to create and modify 3D models, animations, and scenes. It simplifies complex operations and supports real-time export to projects.
PageIndex MCP is an inference-based vectorless RAG system. Through the MCP protocol, it exposes the tree-like index of documents to LLMs, enabling platforms such as Claude to retrieve information from PDF documents through structural reasoning like human experts, without the need for a vector database.
Skillz is an MCP server that can convert Claude-style skills into callable tools for any MCP client to use. It discovers skills in the skill directory, exposes skill descriptions and resources, and can run bundled auxiliary scripts.
A NestJS module for easily exposing AI tools, resources, and prompts via the Model Context Protocol (MCP), supporting multiple transport types and automatic discovery and registration functions.
The CyberEdu MCP Server is the official Model Context Protocol server of the CyberEdu CTF platform. It enables seamless interaction with the cybersecurity training platform by dynamically discovering all public methods of the CyberEduClient and automatically exposing them as MCP tools.
MCP Solver is a server that exposes SAT, SMT, and constraint - solving capabilities to large language models through the Model Context Protocol (MCP).
Axone's MCP server is a lightweight implementation that exposes Axone's capabilities through the standardized Model Context Protocol (MCP), serving as a gateway for AI tools to access the data universe.
The CodeSandbox MCP server exposes CodeSandbox SDK operations through tools for AI agents to perform operations such as sandbox creation and file read and write.
A tool that exposes a secure local Python executor through the MCP protocol, providing code interpreter functionality for LLM applications
A project that exposes AutoHotkey functionality through an MCP server, enabling model interfaces to control Windows automation tasks.
MCP Gateway is a server aggregation tool that connects multiple Model Context Protocol (MCP) servers to a unified gateway. It exposes the tools of all connected servers through search, description, and invocation interfaces, effectively solving the problems of context overflow and cognitive overload caused by too many tool descriptions when connecting to multiple MCP servers.
The LDAP MCP server is a tool that exposes LDAP directory services through the Model Context Protocol (MCP), allowing clients to perform directory searches and create, read, update, and delete operations, and supporting secure connections and read-only mode.
A Python template project for quickly building a Model Context Protocol (MCP) server, which provides registration and exposure functions for tools and tips and supports multiple deployment methods.
A PHP-based tool for quickly building MCP servers, supporting the creation of tools, exposure of resources, and definition of prompts, simplifying LLM context management.
This project is a Mongo MCP server built on Spring Boot AI, used to connect to local MongoDB instances and expose database operations through callable tools, and supports testing through MCP clients.
A backend system for managing Kubernetes resources, retrieving and analyzing logs, and formatting and exporting logs through the MCP protocol.
Figma MCP Bridge is a real-time two-way communication MCP server that bridges the Figma plugin through WebSocket, enabling Claude to read and operate on Figma documents. It supports 62 types of Figma operations, including creating shapes, modifying styles, managing components, and exporting resources.
The MCP server implementation of Foxit PDF API, providing Python and TypeScript versions, exposes more than 35 operations (such as creation, conversion, editing, security, OCR, etc.) of Foxit PDF services as tools available to AI agents.