Python · ★ 23,109 · 🍴 2,274 · 📈 1,833 stars today
Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.
中文介绍 本地优先的代码智能图，为 MCP 和 CLI 提供持久化代码库映射。AI 编码工具只读取关键部分，在代码审查和大仓库上减少上下文负担，提升效率。
Rust · ★ 9,612 · 🍴 1,082 · 📈 568 stars today
The most intelligent agent harness for code
中文介绍 面向代码的智能代理框架，简化代码生成、审查和重构流程，为开发者提供全流程 AI 辅助。
TypeScript · ★ 21,759 · 🍴 2,959 · 📈 1,107 stars today
Never stop coding. Free MIT AI gateway: one endpoint, 268+ providers (50+ free), 500+ models — Claude, GPT, Gemini, Kimi K3, GLM, DeepSeek. Works with Claude Code, Codex, Cursor, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, multimodal, Desktop/PWA
中文介绍 免费 MIT 协议 AI 网关，单一端点接入 268+ 提供商（含 50+ 免费）和 500+ 模型。兼容 Claude Code、Cursor 等编码工具，无需切换工作流。
Python · ★ 40,500 · 🍴 6,722 · 📈 823 stars today
Learn it. Build it. Ship it for others.
中文介绍 从零开始的 AI 工程实战指南，涵盖学习、构建到交付全流程，帮助开发者快速上手并落地 AI 项目。
Shell · ★ 134,689 · 🍴 21,962 · 📈 862 stars today
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
中文介绍 一站式 AI 代理平台，提供多种专业代理（前端、社区、创意等），每个代理具备独立人格和工作流程，可组合完成复杂任务。
Python · ★ 18,728 · 🍴 1,463 · 📈 458 stars today
A Flexible Framework for Experiencing Heterogeneous LLM Inference/Fine-tune Optimizations
中文介绍 灵活的异构 LLM 推理与微调优化框架，支持多种硬件和模型组合，帮助开发者探索不同优化策略并提升性能。
TypeScript · ★ 44,146 · 🍴 5,365 · 📈 821 stars today
The open-source AI voice studio. Clone, dictate, create.
中文介绍 开源 AI 语音工作室，支持语音克隆、听写和创意生成，面向语音内容创作者和开发者，提供一站式语音处理能力。
Python · ★ 28,776 · 🍴 2,743 · 📈 234 stars today
Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.
中文介绍 开源 AI 记忆平台，基于自托管知识图谱引擎，为 AI 代理提供跨会话持久记忆，增强上下文理解和持续学习能力。
Python · ★ 14,208 · 🍴 1,483 · 📈 565 stars today
A feed-forward 3D foundation model for reconstructing scenes from streaming data
中文介绍 前馈式 3D 基础模型，从流式数据实时重建三维场景，适用于机器人、自动驾驶等需要在线环境感知的场景。
TypeScript · ★ 5,833 · 🍴 645 · 📈 939 stars today
Open source alternative to Semrush and Ahrefs
中文介绍 开源的 SEO 分析平台，功能对标 Semrush 和 Ahrefs，提供关键词研究、竞争分析等，适合网站优化和营销团队。
Python · ★ 10,226 · 🍴 1,218 · 📈 410 stars today
Kimi Code CLI is your next CLI agent.
中文介绍 Kimi 推出的 CLI 代理工具，在终端中提供智能编码辅助，支持代码生成、解释和调试，提升命令行开发效率。
Python · ★ 37,075 · 🍴 2,577 · 📈 317 stars today
AI Agent Assistant & development framework that integrates lots of IM platforms, LLMs, plugins and AI feature, and can be your openclaw alternative. ✨
中文介绍 开源 AI 代理框架，集成多 IM 平台、LLM 和插件，支持自定义 AI 助手，适合搭建跨平台聊天机器人及自动化工作流。
Python · ★ 26,525 · 🍴 2,164 · 📈 96 stars today
🚀 The fast, Pythonic way to build MCP servers and clients.
中文介绍 快速、Pythonic 的 MCP（Model Context Protocol）服务器与客户端构建框架，简化 AI 工具与模型间通信，适合 Python 开发者。
TypeScript · ★ 2,524 · 🍴 147 · 📈 689 stars today
The go-to web for your AI coding agent — local-first search, fetch, crawl & research over MCP. No API keys, no cloud, $0/query. Public beta.
中文介绍 本地优先的 AI 编码代理网络工具，通过 MCP 实现搜索、抓取和爬虫，无需 API 密钥和云服务，适合本地开发环境。
Rust · ★ 1,528 · 🍴 40 · 📈 371 stars today
A batteries-included framework for building web apps
中文介绍 基于 Tokio 的全功能 Web 应用框架，提供开箱即用的异步开发体验，适合 Rust 生态快速搭建高性能 Web 服务。
TypeScript · ★ 133,767 · 🍴 7,559 · 📈 240 stars today
Collection of publicly available IPTV channels from all over the world
中文介绍 全球公开 IPTV 频道资源集合，收录数千个直播频道，支持多种播放器，适合电视爱好者和 IPTV 应用开发者。
TypeScript · ★ 4,770 · 🍴 331 · 📈 1,641 stars today
Self-hosted deployment platform
中文介绍 开源自托管部署平台，支持快速发布和管理应用，适合希望掌控基础设施的开发者或小团队。
TypeScript · ★ 1,717 · 🍴 230 · 📈 464 stars today
Free, open-source web app for learning about ontologies and Microsoft Fabric IQ. Explore a catalogue of pre-built ontologies, design your own visually, export as RDF/XML, and share interactive diagrams. Zero backend, fully static.
中文介绍 微软开源的本体学习 Web 应用，支持浏览预置本体、可视化设计、导出 RDF/XML 并分享交互图，适合知识图谱初学者。
C++ · ★ 1,236 · 🍴 31 · 📈 395 stars today
ggml speech-to-text inference for 16+ model families
中文介绍 基于 ggml 的轻量级语音识别推理库，支持 16+ 模型家族，可在边缘设备或本地运行，无需 GPU。
C++ · ★ 9,814 · 🍴 518 · 📈 282 stars today
Very low latency speech to text, intent recognition, and text to speech, for building voice agents and interfaces
中文介绍 极低延迟的语音 AI 套件，集成语音转文字、意图识别和语音合成，专为构建实时语音代理和交互界面设计。
👍 182
A growing gap separates inference context lengths from RL post-training: inference systems are approaching million-token contexts, while post-training workloads often remain at 256K tokens or below and rely on length generalization at deployment. The gap is especially important for AI agents, whose
中文介绍 LongStraw 提出在固定 GPU 预算下将强化学习后训练扩展到 2M token 以上上下文的方法。通过分阶段训练策略和高效的注意力机制，在长上下文推理任务中实现了超越现有方法的性能，且显存开销可控。
👍 155
Recent advances in video understanding have spanned motion, long video, and streaming interaction, driving this field toward real-world applications. Despite this progress, current open-source models remain limited in several ways. They often struggle to generalize across diverse video types, making
中文介绍 VideoChat3 是一个全开源视频多模态大模型，支持高效且通用的视频理解。采用统一的视觉编码器和多任务训练策略，在运动、长视频和流式交互等多种任务上实现优异性能，缩小了开源模型与闭源模型的差距。
👍 113
Skills are a useful abstraction for software agents, turning human and agent experience into reusable procedural knowledge. Yet existing skill libraries are mostly hand-written, text-centric, or derived from agent traces, leaving tutorial videos and other multimodal human resources largely underused
中文介绍 RESOURCE2SKILL 提出从人类创建的多模态资源（如教程视频）中提炼可执行智能体技能的方法。通过将视觉、语言和动作信号对齐，自动构建可复用的技能库，显著降低手工编写成本，并提升技能在未知环境中的迁移能力。
👍 111
Graph retrieval-augmented generation (GraphRAG) enhances large language models with structured knowledge, yet existing systems construct knowledge graphs in a single extraction pass, producing noisy entities and brittle retrieval. RAGU, an open-source modular GraphRAG engine, addresses this by separ
中文介绍 RAGU 是一个开源 GraphRAG 引擎，通过分离知识图构建和检索阶段减少噪音。采用多步检索策略和领域自适应 LLM，在复杂问答任务上准确率提升 15% 以上，检索效率也显著优于单步方法。
👍 63
Recent advances in Tool-Integrated Large Language Models have made web search a core capability of information-seeking agents. However, as interaction histories grow, agents increasingly struggle to track task progress. When search attempts fail to yield useful evidence, current single- and multi-ag
中文介绍 SearchOS-V1 提出多智能体协作框架解决开放域信息搜索中的任务跟踪丢失问题。通过专门的追踪代理和动态重规划机制，在长交互场景下搜索成功率提升 20%，并支持跨任务重用。
👍 55
We present Loopie, the most powerful looped Transformer to date. The Loopie series consists of two Mixture-of-Experts (MoE) models: a 20B-parameter model with 2B active parameters and a 6Bparameter model with 0.6B active parameters. Looped Transformers have long faced a challenge: given an N-fold in
中文介绍 Loopie 系列是目前最强的循环 Transformer，包含两个 MoE 模型（20B/2B 激活和 6B/0.6B 激活）。通过多轮循环推理，在多个 NLP 基准上以更少参数达到甚至超越非循环模型的性能，验证了循环结构作为参数高效扩展的可行性。
👍 54
We present Xiaomi-Robotics-1, a foundational vision-language-action (VLA) model capable of (1) following diverse language instructions to perform a wide range of mobile manipulation tasks in unseen environments out-of-the-box, and (2) efficiently adapting to novel downstream tasks with minimal fine-
中文介绍 Xiaomi-Robotics-1 是基于 10 万小时真实机器人轨迹训练的视觉-语言-动作基础模型。支持零样本跨环境移动操作，并通过少量微调适应新任务，在多场景成功率上达到 85% 以上，展示了大规模真实数据驱动的通用机器人能力。
👍 50
World-action models (WAMs) are emerging as a promising foundation for embodied control: rather than predicting actions alone, they learn representations that couple action generation with future world prediction. This coupling is often viewed as a source of robustness, interpretability, and safety,
中文介绍 BadWAM 揭示世界-动作模型在联合预测世界状态和动作时可能出现「正确想象但错误行动」的病理。通过分析耦合表示中的冲突，提出解耦训练策略，在机器人控制任务中减少 30% 以上的错误动作，同时保持世界预测准确性。
👍 47
Hyper-Connections (HC) expand the residual stream of Transformers into N parallel streams, providing a form of memory scaling beyond model width and depth. Manifold-Constrained HC (mHC) stabilizes this formulation at scale. The large gains from N{=}1 to N{=}4 suggest residual-stream expansion as a p
中文介绍 xHC 扩展超连接以进一步扩大 Transformer 残差流并行分支数（从 N=4 到更大）。在多个规模下，xHC 相比原始 Hyper-Connections 在困惑度和下游任务上获得稳定提升，表明残差流扩展是一种有效的记忆缩放维度。
👍 43
Healthcare spans high-stakes communication, expert reasoning, and workflow execution, yet specialized LLMs that cover these use cases together remain limited. A healthcare model must handle patient consultation, clinical reasoning over text and images, interactive diagnosis, and electronic health re
中文介绍 Cura 1T 是针对医疗场景的专业大模型，覆盖患者咨询、临床推理、交互诊断和电子健康记录处理。在医疗基准上以 1T 参数达到 SOTA，并通过多任务训练和领域指令微调，在对话和推理任务中表现优异。
👍 38
Video generation increasingly relies on keyframe-based workflows, where creators specify a sequence of reference images to guide generation. Although recent models support multi-keyframe conditioning, it remains unclear whether they can faithfully reproduce the prescribed keyframes while maintaining
中文介绍 KeyFrame-Compass 提出首个关键帧条件视频生成的综合评测基准。系统评估模型在保持指定关键帧内容的同时生成平滑视频的能力，发现现有方法在关键帧忠实度和时序一致性上存在明显不足。
👍 33
Multi-reference-to-audio-video (MR2AV) generation aims to generate coherent audio-video content conditioned on multiple references and textual instructions. Existing benchmarks mainly focus on text-driven generation, single-reference subject preservation, or isolated audio-video alignment, leaving t
中文介绍 MultiRef-Compass 针对多参考到音频-视频生成任务构建评测基准。评估模型在同时遵循多个图像参考和文本指令生成音视频内容时的对齐度与一致性，填补了该领域缺少系统化评测的空白。
👍 31
Human cognition does not separate understanding and generation. A teacher at a whiteboard speaks and draws together, each modality reshapes the other. In this paper, we bring this coupled loop to artificial systems. Masked Diffusion Models (MDMs) are ideally suited to this task, yet existing sampler
中文介绍 提出自纠错耦合马尔可夫跳转过程，实现同时的图像理解与生成。利用掩码扩散模型的双向特性，在生成过程中根据理解反馈修正生成结果，在图像修复和编辑任务上质量显著优于独立生成-理解流水线。
👍 30
Learning broad world knowledge directly from raw visual data is a fundamental capability of intelligence. We introduce UniVR, the first investigation into simultaneously learning complex reasoning, fine-grained physical dynamics, and long-term planning from pure visual demonstrations. At its core, U
中文介绍 UniVR 首次探索仅从纯视觉演示中同时学习复杂推理、物理动态和长期规划。通过将视频编码为结构化视觉表示并引入推理链，在视觉问答和规划任务上达到或超越多模态方法，展示了纯视觉世界知识学习的潜力。
👍 27
On-policy distillation is an alternative post-training method in reinforcement learning that alleviates the constraints imposed by reward models by providing token-level supervision from a teacher model. Although on-policy distillation has been studied and applied across various settings, its fundam
中文介绍 On-Policy Delta Distillation 提出基于 token 级教师监督的强化学习后训练新方法。通过仅蒸馏教师与学生的策略差异，在保持性能的同时大幅降低计算开销，在多个 LLM 基准上训练速度提升 2 倍。
👍 26
Large language models (LLMs) are transforming recommender systems from matching co-occurrence patterns in historical behavior toward reasoning about the intent that drives it. RecGPT-V1 pioneered this paradigm on Taobao by centering user understanding, and RecGPT-V2 scaled it via coordinated multi-a
中文介绍 RecGPT-V3 是面向淘宝推荐系统的大语言模型，从行为匹配升级为意图推理。通过多任务协同训练和用户理解中心范式，在点击率、转化率等核心指标上显著超越 V2 版本，展示了 LLM 用于推荐系统的有效路径。
👍 25
Embodied cognition requires agents to connect high-level task reasoning with the physical states to be achieved. We introduce Hy-Embodied-RxBrain, an embodied cognition foundation model with joint language-visual reasoning and imagination. Unlike vision-language models that emphasize scene understan
中文介绍 RxBrain 提出具身认知基础模型，联合语言-视觉推理与想象。不同于强调场景理解的 VLM，该模型通过想象未来物理状态指导动作生成，在机器人操作任务中泛化能力提升 25%，并支持零样本迁移。
👍 24
Code review helps maintain software quality before code integration, but it also imposes a substantial workload on human reviewers. As generative artificial intelligence becomes part of software development, code review is shifting from a primarily human review process toward AI-supported review pro
中文介绍 系统研究不同代生成式 AI 技术（LLM、代码模型等）对代码评审质量的影响。发现 AI 辅助能减轻审查者负担，但在捕获深层逻辑缺陷上仍弱于人类，且不同代际模型在评审建议的准确性和相关性上存在差异。
👍 24
Reinforcement learning has become a standard post-training recipe for large language models, but dense full-parameter updates create two deployment-relevant bottlenecks: suppressed reasoning performance, often reflected by premature saturation of test-time scaling, and interference when consolidatin
中文介绍 Spectral Rewiring 提出基于谱方法对强化学习微调后的 LLM 进行探索、净化与模型合并。通过调整权重矩阵的奇异谱，缓解推理性能饱和并减少多任务干扰，在数学推理和代码生成任务上测试时扩展曲线更优。
👍 20
On-policy distillation (OPD) has become a key paradigm in LLM post-training, yet its training dynamics remain poorly understood. We present a systematic study examining the role, pathologies, and regulations of OPD. We first clarify the role of OPD as an exploration catalyst: it steers the student t
中文介绍 系统研究在线策略蒸馏的训练动力学：其核心角色是「探索催化剂」，引导学生模型探索错误区域。发现病理如蒸馏崩溃，并提出正则化方法加以调控，在多个 LLM 基准上提升训练稳定性和最终性能。
@nifinet · 11.0K 粉丝 · 337.0K 阅 · 518 赞 · 33 转
Earlier this year, Andrej Karpathy (@karpathy) pointed an agent at his own training code and let it run for two days. It ran 700 experiments, kept the 20 that beat the benchmark, and made the model
中文介绍 Andrej Karpathy 让一个 agent 针对自己的训练代码运行两天，完成 700 次实验，仅保留 20 个超越基准的结果。本文介绍如何在 Codex 上构建能自我改进的出站系统。
@CliffordSosin · 16.4K 粉丝 · 197.0K 阅 · 514 赞 · 49 转
Superintelligence arrived. You probably didn't notice, because it turned out to be kind of incremental. Don't believe me? Run the test. Talk to Fable 5 for an hour, then talk to your ten smartest
中文介绍 作者认为超级智能已来但被低估，建议通过对话 Fable 5 来验证其能力。指出智能提升可能只是渐进式的，并不惊艳。
@EXM7777 · 126.7K 粉丝 · 173.6K 阅 · 521 赞 · 51 转
I'm going to show you how to build your first team of AI agents... and how to wire them into a self improving loop inside a shared workspace the team we're building has 5 agents, one for each function
中文介绍 教程：如何构建一个由 5 个功能 agent 组成的团队，并将它们连接到一个共享工作区中的自我改进循环。
@0xCodez · 23.9K 粉丝 · 154.5K 阅 · 510 赞 · 73 转
Most people who try to build a multi-step agent end up with a straight line. Step one, step two, step three - each waiting politely for the last to finish before it starts. 9/10 notice that half those
中文介绍 多步 agent 容易变成串行流水线（90% 情况下步骤冗余），本文提供用 Claude 实现图工程的 14 步路线图，从零到图架构师。
@herrcore · 15.4K 粉丝 · 64.2K 阅 · 514 赞 · 84 转
tl;dr K3 makes us question why we are paying Anthropic. We’ve been using Opus as our go-to model for reverse engineering tasks since the release of 4.5, and with each new release, it feels like we
中文介绍 使用 Kimi K3 进行逆向工程任务，性能超越 Opus，让人质疑 Anthropic 的收费合理性。作者分享了 K3 如何做到“没有 nag”的体验。
@IntuitMachine · 62.3K 粉丝 · 50.7K 阅 · 552 赞 · 47 转
What the shift in AI agent architecture is really about Peter Steinberger just posted nine words that gathered thousands of likes: https://x.com/steipete/status/2078277297791189132 "Are we still talking loops or did we
中文介绍 讨论 AI agent 架构正从“循环工程”转向“图工程”，引用 Peter Steinberger 引发数千赞的观点：我们还在谈循环，还是已经转向图了？
@enginoid · 1.1K 粉丝 · 49.9K 阅 · 522 赞 · 37 转
Linear is probably my best-value subscription at the moment. I pay the sole user fee of $12/mo and the company agents have eaten through ~400 issues in the past month. I use Linear to organize my
中文介绍 作者用 Linear 管理 agent 任务：月费 12 美元，agent 一个月处理约 400 个 issue。分享如何构建类似软件工厂的 agent 工作流。
@therealDeFlock · 82.7K 粉丝 · 12.2K 阅 · 500 赞 · 120 转
"Axxon One supports automatic tracking of people as they move between cameras". "automatic tracking of people"... Ideal for logistics, warehousing, and Safe City projects, this AI detector
中文介绍 介绍 Axxon One：AI 人物检测器，支持跨摄像头自动跟踪，适用于物流、仓储和 Safe City 项目。
This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. Over the weekend, several current and former advisors to President Donald Trump on AI publicly lobbed insults at the country’s leading AI companies. David Sacks,
中文介绍 中国AI模型引发特朗普团队内部分裂，多位现任及前任AI顾问公开批评中国领先AI公司，中美AI竞争加剧。
中文介绍 英伟达发布Cosmos 3 Edge模型，具体技术细节尚未披露。
OpenAI shares lessons from deploying long-running AI models, highlighting new safety risks, observed failures, and improved safeguards through iterative deployment.
中文介绍 OpenAI发布长周期模型安全对齐报告，分享部署经验，指出新安全风险与失败案例，并通过迭代部署改进防护措施。
The next time you apply for a job, AI may screen your résumé before any human sees it. But there’s good reason to question whether AI will judge you fairly. Researchers already know that LLMs pick up human biases from their training data. New research suggests that LLMs can also develop their own bi
中文介绍 新研究显示，AI在招聘中比人类更容易形成偏见，LLM从训练数据习得偏见，可能导致简历筛选不公。
中文介绍 TLDR AI日报：Qwen 3.8发布，Kimi Code CLI上线，Netflix公开其LLM技术栈。
a quiet day
中文介绍 AI新闻简报报道当日较为平静，无明显重大事件。
Sarah Friar, CFO of OpenAI, introduces a practical AI scorecard to measure ROI through useful work, cost per successful task, dependability, and return on compute.
中文介绍 OpenAI CFO Sarah Friar推出AI时代计分卡，通过有用工作、每成功任务成本、可靠性和计算回报衡量AI投资回报率。
Every morning, airline dispatchers, grid operators, and farmers around the world make decisions based on the same thing: a weather forecast. While these forecasts are something that most people glance at for two seconds, weather predictions influence major strategic decisions in many industries, wit
中文介绍 天气数据遭破坏风险上升，天气预报影响航空调度、电网运营和农业决策，数据安全至关重要。
**Moonshot's Kimi K3 release** has sparked a reassessment of **Chinese open-weight models**' proximity to the frontier, with strong performance in coding, agentic tasks, and long-horizon knowledge work. The strategic focus has shifted from a "compute moat" to an "efficiency stack" involving **MoE ro
中文介绍 Moonshot发布Kimi K3，引发对中国开源模型前沿能力重新评估，在编程、智能体和长知识任务中表现强劲，战略重心从计算护城河转向效率堆栈。
a great week for open models continues.
中文介绍 Kimi K3（2.8T-A50B）发布，成为史上最大开源模型，性能达Opus 4.8级别，定价与Sonnet 5相当。
中文介绍 TLDR AI日报：Kimi K3发布，Gemini 3.5推迟，ARC-AGI 3被攻克。
Learn how OpenAI is making ChatGPT safer for teens with age-appropriate protections, learning tools, parental controls, and expert partnerships.
中文介绍 OpenAI阐述青少年应获得安全AI访问权，为ChatGPT添加年龄适宜保护、学习工具、家长控制及专家合作。
Lila is betting that science, not the internet, is the last untapped source of training data. We went to find out what that actually looks like in a room full of robots.
中文介绍 Lila Sciences的Andy Beam和Rafa Gómez-Bombarelli认为未来实验室应像数据中心，科学是最后未开发训练数据，机器人实验室正实现此愿景。
中文介绍 Dharma AI博客探讨新模型保持相同优势，具体内容未详。
Google DeepMind and Isomorphic Labs are sharing our joint approach to bioresilience and AI models.
中文介绍 Google DeepMind与Isomorphic Labs发布联合生物韧性方法，分享AI模型在生物领域应用的最新进展。
美股方面，标普500 ETF（SPY）和纳斯达克100 ETF（QQQ）均出现MACD死叉信号，RSI分别处于47.5和42.3的弱势区域，短期技术面承压。加密市场受恐慌贪婪指数仅25（极度恐慌）影响，总市值2.31万亿美元、BTC主导率56.5%、ETH占比9.9%，比特币和以太坊RSI分别为56.6和62.2，虽MACD金叉但价格仍低于200日均线，反弹动能有限。中概股表现分化，阿里巴巴和京东近5日涨幅超5%，京东RSI达70.3进入超买状态，拼多多虽呈空头排列但RSI中性。商品外汇方面，黄金期货RSI 40.1且均线空头排列，走势疲弱；WTI原油RSI 58.2但价格低于50日均线，方向不明；美元指数和10Y美债收益率均呈多头排列并接近52周高位，反映利率预期偏鹰。整体而言，各资产类之间缺乏一致性向上动能，市场处于结构性调整阶段。
当前价格326.59美元，高于20日（306.95）与50日（303.43）均线，形成多头排列。RSI 14为63.9，处于中性偏强区域。MACD值8.88，信号线6.16，呈金叉且柱线持续放量。近5日上涨2.92%，距52周高仅差2.51%，显示上行趋势完好。
价格369.57低于20日（396.31）、50日（409.22）及200日（416.67）均线，呈空头排列。RSI 14为39.2，处于弱势区域。MACD死叉（-5.50 vs -2.68），柱线为负。近1日下跌2.96%，近5日下跌6.38%，空头动能明显。
价格30.6高于20日（27.33）、50日（28.91）及200日（29.59）均线，均线呈多头排列；但RSI 14达70.3，进入超买区域，暗示短期回调风险。近5日上涨5.96%但动能受超买信号压制，技术面方向待定。
VIX 恐慌指数
10Y 美债收益率 (%)
美元指数 DXY
S&P 500 ETF
Nasdaq 100 ETF
Apple
Microsoft
Nvidia
Alphabet
Tesla
Meta
Bitcoin
Ethereum
Solana
阿里巴巴 (BABA)
拼多多 (PDD)
京东 (JD)
腾讯控股 (0700.HK)
黄金期货
WTI 原油期货
美元 / 人民币
过去走势不代表未来表现，技术指标仅反映历史数据计算的统计特征。本报告仅供指标解读参考，不构成任何投资建议，请读者独立判断并注意风险。
The duties mark a major escalation in trade tensions between the North American neighbours.
中文摘要 美国总统特朗普对加拿大进口商品加征50%关税，标志着北美邻国间贸易紧张局势重大升级。
Airline says communication ‘could have been better’ after about 250 passengers waited hours for flight to Denpasar to leave. Follow today’s news live Get our breaking news email, free app or daily news podcast Alpine ski resorts in New South Wales and Victoria saw climate records shattered at the we
中文摘要 澳航就约250名乘客在悉尼停机坪等待四小时道歉，称沟通本可更好。议员巴纳比·乔伊斯在7.30采访中对狗发火。
Young people angry about frequent leaks of college entrance tests present one of the biggest challenges in years to Prime Minister Narendra Modi.
中文摘要 印度CJP党（蟑螂党）因大学入学考试频繁泄露，领导抗议游行前往议会，对总理莫迪构成近年来最大挑战。
US Central Command says latest attack aimed at degrading Iranian military capabilities after president threatens country over deaths of US troops US launches 10th round of strikes as Iran says it is in ‘full-scale war’ The price of oil jumped in early Asian trade on Monday after the US carried out i
中文摘要 美军对伊朗发动新一轮打击以削弱其军事能力，特朗普威胁伊朗将付出代价。伊朗称处于全面战争状态，油价在亚洲早盘上涨。
Police say a possible political motive is under investigation in the killing of the Reform UK figure.
中文摘要 嫌疑人被控杀害英国改革党人物安·威德科姆，警方称正调查可能存在的政治动机。
中文摘要 暂无详细信息。
Britain has a new Prime Minister: Andy Burnham. The former Mayor of Manchester must now unite a divided nation.
中文摘要 英国新任首相安迪·伯纳姆（前曼彻斯特市长）上任，面临团结分裂国家的重大挑战。
New report denounces Havana and claims ‘campaign of subversion’ has infiltrated highest levels of US government As part of the Trump administration’s continuing attack on left-leaning groups, the US state department has listed several activists and organizations as “front groups and fellow travelers
中文摘要 美国国务院发布报告，将多个活动家和组织列为针对古巴的“前线团体和同路人”，称其“颠覆运动”已渗透美国政府高层。
Chaos outside India's parliament as protesters tried to storm it; satirical Cockroach Janta Party gains traction.
中文摘要 印度议会外爆发混乱，抗议者试图冲击议会，讽刺性的“蟑螂党”活动人士获得关注。
Turmoil likely as Trump officials say Canada unfairly discriminated against US autos, alcohol and dairy products Donald Trump is imposing 50% tariffs on most Canadian goods in response to the country retaliating against previous US tariffs, the White House announced Monday, declaring Canada has unfa
中文摘要 白宫宣布特朗普将对大部分加拿大商品征收50%关税，报复加拿大此前对美关税措施，预计引发经济动荡。
30 days into the US-Iran MoU the worst fighting in months has erupted threatening to unravel the ceasefire.
中文摘要 美伊备忘录签署30天后，最激烈战斗爆发，威胁破坏停火。
A video showing the final moments of Moroccan-born Abderrahim Fakir as he was detained by police has sparked protests.
中文摘要 摩洛哥裔男子阿卜杜拉希姆·法基尔在意大利被警方拘留期间死亡，相关视频引发抗议。
New York’s mayor, Zohran Mamdani, told The New York Times, he was looking into whether he could arrest Prime Minister Benjamin Netanyahu when he comes to New York for the U.N. General Assembly.
中文摘要 纽约市长佐兰·马姆达尼表示正研究能否在以色列总理内塔尼亚胡出席联合国大会时将其逮捕，特朗普称不会发生。
A 28-year-old man has been charged in connection with the death of Ann Widdecombe, a right-wing figure whose body was found in her home this month.
中文摘要 英国检察官对右翼人物安·威德科姆之死提出谋杀指控，28岁男子被控，其尸体本月在家中被发现。
In a new book published on Tuesday, Leo said that a race to nuclear de-escalation has been “obscured by a race to armament that today frightens and terrifies.”
中文摘要 教宗利奥在新书中呼吁核裁军，称核降级竞赛已被“军备竞赛所掩盖”，令人恐惧。
The duties mark a major escalation in trade tensions between the North American neighbours.
中文摘要 特朗普对加拿大进口商品加征50%关税，此举标志着北美邻国间贸易紧张局势显著升级。
Australia’s stock market has emerged as an unlikely haven from the artificial intelligence-driven volatility that has roiled Asian equities in recent months.
中文摘要 澳大利亚股市近期成为人工智能驱动波动中的避险地，尽管亚洲股市整体受AI交易波动影响，但澳股表现相对稳定。
Gold was steady as traders monitored a raft of developments in the Middle East conflict for clues on the energy price impact on inflation.
中文摘要 金价持稳，交易员关注中东冲突进展，以判断能源价格对通胀的影响。
Firms have to balance hi-tech and age-old skills when training up young workers.
中文摘要 企业培训年轻员工时需平衡高科技与传统技能，既掌握数字技术又传承古老工艺。
Buying on social media live streams is fast becoming regular consumer behaviour for many.
中文摘要 直播购物日益普及，有卖家在TikTok上单日销售额达10万英镑，社交电商正成为常规消费行为。
New Zealand’s annual inflation accelerated to the fastest pace in more than two years, fanned by soaring fuel prices and bolstering the case for the Reserve Bank to execute back-to-back interest-rate increases.
中文摘要 新西兰年通胀率加速至4.1%，创两年多最快增速，受燃料价格飙升推动，强化了纽储行连续加息的预期。
Analysts are maintaining bullish views on Kioxia Holdings Inc. despite a deepening share price correction, saying the stock is likely to resume its upward trajectory once supply-demand conditions improve.
中文摘要 尽管股价回调加深，分析师仍对铠侠控股持看涨观点，目标价高出当前市场价130%，预计供需改善后股价将恢复上涨。
The bankrupt $1.2 billion Oceanwide Plaza project in Los Angeles, dubbed the Graffiti Towers after taggers spray-painted the abandoned complex, has been cleared for sale and possible completion more than seven years after the original Chinese developer ran out of money.
中文摘要 洛杉矶价值12亿美元的Oceanwide Plaza项目（因涂鸦被称为「涂鸦塔」）获准退出破产，该综合体由原中国开发商资金断裂后闲置超7年，现已可出售并可能完工。
Oil steadied after a two-day gain as the US and Iran continued hostilities in the Middle East, while Yemen’s Houthi militants threatened to blockade Saudi Arabia’s maritime traffic in the Red Sea.
中文摘要 油价在两日上涨后持稳，美伊在中东持续冲突，也门胡塞武装威胁封锁红海航运，影响沙特海运。
Oil held near its highest close since mid-June, weighing on stocks and Treasuries, as an escalation in Middle East hostilities rekindled worries about inflation.
中文摘要 油价接近6月中旬以来最高收盘价，中东冲突升级重燃通胀担忧，拖累股市和国债。
The BBC hears from people about the risks and rewards of investing in technology shares.
中文摘要 BBC采访了一些20多岁年轻人，他们大举投资科技股，面临风险与回报，反映了年轻一代的投资热情。
Every tricky decision and trade off is being watched by the country and the markets at the same time
中文摘要 伯纳姆提出宏大计划，但每个决策和权衡都受到全国和市场关注，成本问题待解。
10 回复 · 程序员 节点
6 回复 · Apple 节点
11 回复 · 程序员 节点
19 回复 · 程序员 节点
20 回复 · 程序员 节点
16 回复 · Apple 节点
7 回复 · Apple 节点
14 回复 · Apple 节点
20 回复 · Apple 节点
33 回复 · Linux 节点
该源今日无内容。