We're releasing HeyGen Video, built for businesses that need production-quality video without production-level costs. Pricing starts at $0.01/s through October (50% off) Built on @Minimax_AI H3, post-trained by HeyGen. Learn more: https://developers.heygen.com/heygen-video-1.0-catalog
#产品更新
#产品更新
今日 73 条
MiniMax (official)@MiniMax_AIAI 评分4848引用HeyGen@HeyGen
eric zakariasson@ericzakariassonAI 评分5353引用Grok Bot@botGrok Bot is now more powerful for building software. Bots can hand off coding tasks to Cursor, manage your PRs with GitHub and Origin plugins, and share video demos of what they build.
Aravind Srinivas@AravSrinivas精选AI 评分6969引用Perplexity@perplexity_aiComputer now works in email. Send, forward, or cc computer@perplexity.com on any thread. Every email task runs as a normal session in Computer, viewable on web and mobile, with the same audit trail as any task in the app.
推荐理由:Perplexity 开放无需账号的邮箱入口并限时免费,读者可以据此评估把任务转给 Computer 智能体执行的可行性。
OpenAI Developers@OpenAIDevsAI 评分3838可分享的个人资料将你的 Sites 和插件汇集在 ChatGPT 中,让他人能够找到并复用你的作品。 做好了?就秀出来。

🚨 AI News | TestingCatalog@testingcatalogAI 评分3131
Runway@runwaymlAI 评分3131
Manus@ManusAIAI 评分3232用 Manus Game Dev 把一个想法变成好玩的 3D 赛车游戏。🏎️ 赛车呢?由我们的合作伙伴 @tripoai 生成的 3D 模型。

The Decoder:AI News(RSS)AI 评分6868 Google 在 Gemini 中全球推出 Skills,取代 Gems
Google 在 Gemini 聊天中全球推出 Skills,一种可由 AI 协助编写和完善的任务详细提示词格式,取代 Gems。用户将常用指令保存为 Skill,输入 "/" 加名称调用,Gemini 还能从历史对话生成 Skills、自动运行匹配的 Skill、链式组合多个 Skills,并支持文本文档、PDF 和图片作参考材料。
OpenClaw🦞@openclawAI 评分4040
Rohan Paul@rohanpaul_aiAI 评分4545引用Christian Pickett@chrisspickettIntroducing OpenAI’s Dots x Orthogonal Try it out: https://orthogonal.com Ask for a pipeline before bed. Wake up to verified leads, cited sources and the first replies. Your always-on dots now have every tool they need, from company and people data to email.
OpenRouter@OpenRouterAI 评分3939
TypeSafe AI@typesafeaiAI 评分2828给后排的人: • 用 Jev 对销售数据做生产级重排序 • 快 20 倍 • 便宜 10 倍 • 准确率高 12% 你们懂了吗?
引用Rox@rox_aiWe used Jev to retrieve sales data 20x faster, 10x cheaper, and 12% more accurate than GPT-5 Mini. Every time a Rox agent answers a query, it pulls from relevant transcripts, emails, CRM notes, news, and documents. We benchmarked two ways of retrieving them: LLM-based reranking Jev classification The results speak for themselves. Jev was faster, cheaper, and more accurate. Revenue agents require thousands of points of context to understand accounts, chart relationships, and execute on sales. This research will help us serve that context with frontier-level performance for our customers.
elvis@omarsar0AI 评分5050
引用Vivix.AI@VivixLabs_HQCreate your own character. Call the shots. Live. Introducing A1 Playground + Agent: build a live AI character that talks, moves, and plays with you. Create yours: https://platform.vivix.ai/vivix-a1-model/ Their look. Their personality. Their voice. Your choice. Describe who you have in mind, upload a picture, and let our Agent bring them to life. And they do a lot more than talk. 💃 Full-body performance. From air guitar to victory dances. 🚶 Move around the scene. Walk and talk. Pace while making a point. 🛍️ Interact with objects. Pick up products, show them off, and pitch them. 🎮 Play web games together. Deal cards, keep score, and play along. Solve a mystery with Sherlock Holmes, then interrupt him for a dance break. Take a walk with Socrates. Hand him a protein shake and debate the meaning of gains. Give Snape a bottle of shampoo. Ask for the most enthusiastic sales pitch of his life. Who will you bring to life?
🚨 AI News | TestingCatalog@testingcatalogAI 评分4646Creatify 推出广告视频模型 Boreal-H3 及由 Claude Opus 5.5 驱动的 Ad Agent,Boreal-H3 基于 MiniMax H3 用真实广告项目后训练。
引用Creatify Labs@Creatify_LabsIntroducing Boreal-H3 — a video model built for ads and our next step toward recursive self-improvement in video generation. A good-looking video isn’t enough. The product has to stay the same. The actor has to stay the same. The label has to be right. And the action in the brief actually has to happen. So we post-trained MiniMax H3 specifically for advertising. But this isn’t a one-off SFT or LoRA fine-tune. We built a closed-loop system that learns what to improve next. Human-calibrated evaluation diagnoses failures and guides the next intervention: targeted data collection, reinforcement learning, or inference optimization. When the feedback is unreliable, we revise the evaluator or reward—not just the generator. Every experiment feeds into shared memory, informing the next training decision. The model improves, and so does the process that produces its successor. The results: → 85.3% reference fidelity — highest among the frontier video generation models we evaluated → Brief success: 28% → 50% → Identity match: 83% → 94% → Visible defects per clip: down 70% → Generation time and estimated cost: down 20% Boreal-H3 doesn’t just make better-looking video. It makes more usable ads. Credit to the @MiniMax_AI team for the foundation we’re building on. This launch is a checkpoint, not the finish line. We’re building more than a better video model. We’re building a system that learns how to make the next one better.
OpenClaw🦞@openclawAI 评分88
Runway@runwaymlAI 评分4040
Demis Hassabis@demishassabisAI 评分6363引用Pushmeet Kohli@pushmeetVery happy to announce that our team @GoogleDeepmind has pushed the boundaries of generative biology, achieving the successful synthesis of AI-designed proteins that are both functional and watermarked. This proof-of-concept watermarking of the building blocks of life is enabled by SynthID Bio, our new protein watermarking method. It is designed to safeguard the new era of AI-powered generative biology and strengthen global biosecurity. You can read my thoughts here on why watermarking AI-designed proteins is an important research breakthrough: https://x.com/pushmeet/status/2105314763148321102
Elon Musk@elonmuskAI 评分4747引用matt palmer@mattypWhat's new in Grok Bot 0:19 Team bots - create bots shared with your team members 1:10 Finances - search and manage your finances with @plaid 1:32 Voice calls - @bot can search chats during calls and gracefully handles interruptions Like this for a special bot animation on @X!
Alexandr Wang@alexandr_wangAI 评分1515引用Brian Sozzi@BrianSozziMuse still crushing the charts:
Google Cloud:Databases(RSS)AI 评分4848 Spanner Omni 正式 GA:可部署于任何环境的分布式多模型数据库
Google 的 Spanner Omni 正式 GA,这是 Spanner 的可部署于任何环境的版本,支持本地数据中心和其他云上的生产负载。它提供与全托管 Spanner 相同的核心能力,涵盖 SQL、图、键值、全文搜索、向量搜索和列式分析引擎,并支持 MCP Toolbox 接入智能体系统。
Google AI:DEV 作者专属(RSS)AI 评分5959 作者发布 Promptline,用一键热键把常用提示词变成可复用词汇
作者发布开源工具 Promptline(MIT,Windows 10/11,macOS 开发中),把反复手打的提示词命名为可复用词汇,通过 Ctrl+Alt+V 热键把选中提示词连同剪贴板内容粘贴到 Claude Code、Codex、Cursor 等窗口。
Google Blog:AI(RSS)精选AI 评分6666 Gemini 全球上线 skills 功能,将逐步取代 Gems
Google 将 skills 功能直接推向 Gemini 聊天的全球用户,并将在未来几周内面向 Google Workspace 商业、企业、非营利和教育客户开放。
推荐理由:官方公告说明了 skills 的使用方式、可用范围和 Gems 退出时间表,用户可据此安排迁移。
Rohan Paul@rohanpaul_aiAI 评分4242引用Max Derevy@MaxDerevyWe built an AI that saves your ass. Meet Lucas. It’s been meaning to text you. It lives in your texts, learns your life, and prompts itself. It books, pays, orders and checks you in. Usually before you’ve thought of it. Overnight, it goes exploring for things your life could use. Text Lucas today on iMessage and WhatsApp http://www.meetlucas.ai
Suno@sunoAI 评分1414对 @banoffeemusic 来说,创新是本能。一首歌直到让她自己感到惊喜才算完成。 看她如何用 v6 汲取意想不到的影响、融合流派,把她的声音带到新的地方。 做打动你的东西。

Krea@krea_aiAI 评分3535Ideogram 4.5 现已登陆 Krea。 反复编辑通常会导致伪影、像素偏移和色彩漂移——这个模型正是为解决可重复编辑而打造。 在下方试试 👇

Every:最新文章(网页)AI 评分4040 Every 推出 Slack 智能体 @Every,把周报和跟进等重复工作变成 Slack 工作流
Every 上线 Slack 智能体 @Every(beta 版),可将周报、跟进等重复性工作转化为在 Slack 中运行的工作流。用户只需在频道中 @Every 下达指令,例如生成周一发货回顾、追查昨日演示的未跟进事项,或根据 Figma 文件起草本周社交内容包。该智能体由大语言模型驱动,官方提示可能生成不准确或不完整的回复,建议在据此行动前先审核其工作。
TechCrunch:AI(RSS)AI 评分4848 Instinct 推出 Instinct Selections 商品推荐,部分用户反感
AI 智能体 Instinct 上线 Instinct Selections,向用户推送餐饮、旅行、购物等个性化商品推荐,创始人 Noah Shinn 称与本地厨师、设计师、旅行向导等合作引入“人类品味”。
TechCrunch:AI(RSS)AI 评分5151 DoorDash 推出可通过 Apple Messages 发短信点餐的 AI 智能体
DoorDash 周三宣布推出文本点餐 AI 智能体,用户可通过 Apple Messages 发送提示词下单。该智能体能理解“点我常点的”等指令、推荐本地餐厅并生成购物车,还可发送食物照片,处理团体订单中的混合饮食偏好;美国用户可加入等待名单试用。此外 DoorDash 将在北加州与部分餐厅测试配送无人机。
Bloomberg:Technology(RSS)AI 评分3737 CoreWeave CEO 谈 Forge 开发工具与算力需求
CoreWeave 在旧金山首届大型用户大会 CoreWeave Fully Connected 上发布 Forge 开发工具套件,帮助客户训练和改进 AI 模型,意在摆脱“neocloud”标签、拓展产品线。CEO Michael Intrator 称,使用其基础设施的客户正系统性地向技术栈上层迁移,寻求针对 AI 特有问题的云解决方案。
Transluce(网页)AI 评分2828 Transluce 发布 Docent 与 Monitor:面向 AI 智能体行为分析与可观测性的工具
Transluce 推出 Docent 研究预览版,一个用于分析并干预 AI 智能体行为的系统,其技术报告展示了如何用它调查 o3 的真实性。同时上线的 Monitor 是 AI 驱动的可观测性界面,源码与神经元描述数据库已在 GitHub 公开。Transluce 是位于旧金山的独立非营利研究实验室,致力于构建可扩展 AI 监督的公共技术栈。
Microsoft Research@MSFTResearchAI 评分3333极端太空天气事件可能损坏地球电力系统,并降低GPS精度和卫星运行能力。一套新的机器学习系统能在风暴抵达前30-60分钟预测可能发生损坏的位置:https://msft.it/6019aYXwh

NVIDIA AI@NVIDIAAIAI 评分2222构建边缘 AI 的新方式:在 NVIDIA Jetson 上进行智能体开发 https://x.com/i/broadcasts/1yJAPwpzakzGb
Google Gemini@GeminiAppAI 评分4343跳过重复的提示词,用 Gemini 中的技能直接获得结果。 技能是为那些你反复执行的特定任务保存的指令集。把技能想象成一个快捷方式,随时待命。🧵

xAI:News(网页)AI 评分5858 xAI 发布 Grok Speech to Text 与 Text to Speech 独立 API
xAI 发布 Grok Speech to Text 和 Grok Text to Speech 两个独立音频 API,基于驱动 Grok Voice、Tesla 车辆和 Starlink 客服的同一套技术栈。
xAI:News(网页)精选AI 评分6262 xAI 推出 Custom Voices,可克隆声音用于 Grok TTS 和 Voice Agent API
xAI 推出 Custom Voices,录制约一分钟自然语音即可在两分钟内克隆声音,并可在 Grok Text to Speech 和 Voice Agent API 中使用。自定义声音继承全部 TTS 能力(speech tags、多语言输出、REST 与 WebSocket 流式),不额外收费。声音需通过口令朗读和说话人嵌入比对的两阶段验证,无法从已有录音克隆,也无法克隆他人声音。
推荐理由:xAI 官方说明克隆流程、验证机制和 API 覆盖范围,读者可据此评估在语音应用中的可用性。
xAI:News(网页)AI 评分4444 xAI 为 Grok Imagine API 推出 Quality Mode,提升图像生成真实感与文字渲染
xAI 在 Grok Imagine API 上线 Quality Mode,面向企业开发者和团队提供图像生成与编辑能力,主打更高真实感、更强文字渲染和更好的创作控制。该模式适用于产品可视化、营销素材、UGC 风格内容等场景,并可结合视频生成能力制作社交媒体视频、产品展示和广告素材。
xAI:News(网页)精选AI 评分7070 Grok 推出 Connectors 深度集成,并支持自定义 MCP 服务器
xAI 发布 Connectors,把常用应用深度接入 Grok,现已在 Grok Web、iOS 和 Android 上线。
推荐理由:官方公告列出了各连接器的具体读写能力和入口路径,读者可据此评估哪些工具接入 Grok 后能省去复制粘贴。
xAI:News(网页)AI 评分4545 xAI 将 Grok 接入 Nous Research 开源 Hermes Agent
xAI 宣布 Grok 订阅可直接在 Nous Research 的开源自改进智能体 Hermes Agent 中使用,所有订阅档位均可接入。接入后可用 Grok 4.3 进行文本对话与高级推理、Grok Text-to-Speech 语音回复,以及 Grok Imagine 生成图像和视频。
xAI:News(网页)精选AI 评分6868 xAI 为 Grok 推出 Skills 功能
xAI 推出 Skills 功能,教 Grok 一次即可在所有对话中记住偏好和工作流,内置 Word、演示文稿、表格、PDF 和 Skill Creator 等技能,无需设置。用户可通过对话或上传文件自定义技能,自己的版本优先于内置版本。Skills 已在 Grok 4.3 上的 grok.com、iOS 和 Android 上线。
推荐理由:官方公告写明了 Skills 的内置能力、自定义方式和可用平台,读者可据此判断是否纳入自己的工作流。
xAI:News(网页)AI 评分4242 xAI 支持在 OpenClaw 中使用 Grok,SuperGrok 与 X Premium 订阅可直接登录
xAI 宣布即日起 SuperGrok 或 X Premium 订阅用户可在开源、本地优先的智能体 OpenClaw 中登录使用 Grok 模型,所有订阅档位均可接入。