NAll news

All news

Thursday124 items

  1. Google Developers BlogFirst-partyProduct release
    AI score52

    Google API Gateway supports streaming and enters public preview

    Google announced that streaming support for API Gateway has entered public preview. SSE, chunked transfer, WebSocket, and gRPC streaming endpoints can share the same gateway and policies with existing APIs, with no changes required to API specifications. Streams can remain open for up to one hour; during the preview period, this must be enabled when creating a gateway, and existing gateways cannot be converted in place. Combined with model routing, clients can use a single OpenAI-compatible endpoint to route to Gemini, Claude, and OpenAI gpt-oss models based on the model field in the request and return streamed responses. The gateway's default domain name has changed to a fixed format composed of the gateway ID, project number, and region, which can be written into the configuration before creation.

    Streaming feature in API Gateway _ Header

    #Product updates#English text is being prepared.#English text is being prepared.

  2. 36Kr · AIProduct release
    AI score72

    Anthropic Releases Research Preview of Personal AI Agent Claude Duet

    On October 21, Anthropic released the personal AI agent Claude Duet and launched a research preview, initially available to Claude Pro and Max individual subscribers in the United States. The product is based on the Opus 5.5 model and can connect to more than 6,000 tools and services through connectors and plugins in the Claude Marketplace. It also has a cloud-based virtual machine and browser environment, allowing users to continue executing tasks after they exit the app. Anthropic said that sensitive actions such as sending messages, making payments, signing contracts, or deleting data still require user confirmation, and that background work consumes the existing Claude subscription quota.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  3. ThurrottProduct release
    AI score60

    1Password Launches Credential Management Platform 1Password for Agents for AI Agents

    1Password has released 1Password for Agents, allowing supported AI agents to access credentials in a user's personal 1Password vault in order to complete tasks on their behalf. At launch, the platform supports agents such as Anthropic Claude Duet, Meta Muse, OpenAI ChatGPT, and Dots. The agents themselves cannot see the user's credentials; after the user approves the login, 1Password fills in the credentials directly without sending them to the agent. 1Password said that partners including Browserbase, Browser Use, Perplexity, and Ramp will also add support in succession.

    Same story · 13 reports

    #Product updates#Agents#English text is being prepared.

  4. 9to5GoogleProduct release
    AI score68

    1Password Launches 1Password for Agents, Letting AI Agents Complete Logins Without Seeing Passwords

    1Password has launched 1Password for Agents, connecting its password vault to AI agents: when an agent reaches a login form, it authenticates to 1Password using a digital key; after the user confirms which credentials are needed and what they are for, 1Password fills in the form directly—credentials are never handed to the agent and are removed once login succeeds, so the agent cannot read or reuse them. The integration currently supports only Meta's Muse and Grok Bot, with OpenAI Dots and Anthropic Claude Duet to follow.

    Same story · 13 reports

    #Product updates#Agents#English text is being prepared.

  5. SiliconANGLEProduct release
    AI score52

    Action State Launches Open Source Tool to Verify AI Agent Interaction Records

    Action State Group has released an open source toolset for enterprises to independently verify records of AI agent behavior, including EvidenceBook, which preserves tamper-proof records, Terms Packs, which evaluate agent behavior, and Countersign, an independent verification service. EvidenceBook is released under the Apache 2.0 license and uses a signed Agent Action Capsules format and cryptographic checkpoints; verification can be performed offline and is free. Founder Steven Mih said the tool can verify that reported events actually occurred, but cannot confirm whether agreements were fulfilled, and Countersign does not certify the authenticity of results or the integrity of evidence.

    #Product updates#Agents#Open-source ecosystem

  6. SiliconANGLEProduct release
    AI score61

    1Password launches 1Password for Agents, letting personal AI agents log in without touching passwords

    1Password has released 1Password for Agents, a login platform for personal AI agents, in which agents do not see passwords when logging in. In the flow, an agent identifies itself to 1Password via a unique digital key; after the user confirms on their own device which account is to be accessed and for what purpose, 1Password itself fills the password into the page. OpenAI has already integrated ChatGPT Work and the always-on Dots agent; Meta's Muse, SpaceXSI's Grok Bot and Town.com have also signed on, and Anthropic's Claude Duet will be integrated later; another 9 companies are developing integrations, including Perplexity and Browserbase.

    Same story · 13 reports

    #Product updates#Agents#English text is being prepared.

  7. SiliconANGLEProduct release
    AI score56

    Thomson Reuters adds AI agents to its tax and audit software CoCounsel

    Thomson Reuters has released a rebuilt version of the CoCounsel Tax, Audit & Accounting platform, expanding from answering questions to planning tasks, performing calculations, and generating documents for review, and it is now available to all U.S. customers. The platform uses Checkpoint's more than 12 million primary and analytical documents as its research library, supports multiple AI models, can break down complex requests and invoke tools, and provides citations, assumptions, and calculation steps for professionals to review. In a demonstration, the system used a commercial loan agreement and a spreadsheet template to generate an amortization schedule for a $5 million loan, and used a Python routine to recalculate the results as a check; the platform supports MCP and can connect to third-party AI tools such as Claude and Copilot.

    #Product updates#Agents#English text is being prepared.

  8. SiliconANGLEProduct release
    AI score72

    Anthropic launches proactive AI agent Claude Duet

    Anthropic has launched Claude Duet, a proactive AI agent for professionals and small business owners, currently in research preview. Every morning it pushes a briefing of important items from connected apps, checks messages only when attention is needed, and can schedule meetings, build dashboards, generate financial or sales forecasts, and prototype web apps. Duet is integrated into Claude and can directly call existing memory, skills, and apps without additional setup; Anthropic says it follows the trusted agent framework, with automatic classifiers that check for sensitive information leaks and confirm with users before important actions. This version may still misread conversation threads or misjudge urgency, and it cannot show its browser.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  9. The New StackProduct release
    AI score66

    Google opens the infrastructure behind Gemini agents as a cloud service, Agent Compute Engine

    Google announced that it will open a preview of Agent Compute Engine to Google Cloud users later this year, providing each AI agent with a hardware-isolated microVM that balances container startup speed with sandbox security. The service was originally internal infrastructure for Gemini, Antigravity, and Google's model training pipelines; when an agent is idle, it checkpoints the microVM to free up resources. Google is also offering an open-source runtime, Agent Substrate, which can run on any Kubernetes cluster; the company says that in a benchmark of 200,000 agents it achieved 500 activations per second, with suspend and resume taking under half a second.

    Same story · 6 reports

    #Product updates#Agents#English text is being prepared.

  10. The New StackProduct release
    AI score67

    Google open-sources TorchTPU, AWS launches TorchNeuron public beta

    At PyTorch Conference North America 2026, Google open-sourced TorchTPU, a native PyTorch stack for TPU, while AWS released the open-source backend TorchNeuron public beta for Trainium; both are already available on GitHub. TorchTPU was developed by Google in collaboration with Meta's PyTorch team, and its runtime no longer depends on JAX, so existing workloads can be migrated with only minor code changes; TorchNeuron supports Trainium2 and Trainium3 instances, and migration only requires switching to .to('neuron') for eager execution. The PyTorch side points out that TorchTPU's default Skinny Collectives scheme suffers roughly 30–40% end-to-end performance loss on distributed workloads compared with optimized schemes, and achieving peak performance still requires Pallas kernels and production-grade stacks such as vLLM-TorchTPU and SGLang-TorchTPU.

    Same story · 2 reports

    #Product updates#Open-source ecosystem#English text is being prepared.

  11. NVIDIA BlogFirst-partyProduct release
    AI score64

    NVIDIA launches Minerva AIOps platform, delivering end-to-end observability for AI factories

    NVIDIA has launched Minerva, an AIOps platform for AI factories that collects telemetry from infrastructure and AI workloads, then uses purpose-built AI models to turn that data into actionable insights. The platform builds a dynamic knowledge graph that maps GPUs, SuperNICs, switches, cables, and the logic of the jobs running on them, analyzing billions of data points per hour to learn normal behavior, detect unseen anomalies, and predict failures. Minerva includes a built-in AI assistant and can connect to external agentic tools, supporting natural-language troubleshooting of cluster health with remediation recommendations. IREN, IBM, Jio, Lambda, Tesla, and others plan to deploy it, while Dell, Lenovo, and Supermicro will incorporate it into their AI factory operations portfolios.

    Same story · 2 reports

    #Product updates#English text is being prepared.#Agents

  12. TripoFirst-partyProduct release
    AI score55

    Tripo Studio rebuilds rigging and animation features, supports text-to-motion generation

    Tripo has rebuilt the rigging and animation features in Tripo Studio, aimed at game-ready motion production. New features include full-body and hand skeleton rigging, 5 engine-ready skeleton presets, Text to Motion that converts prompts into 1–10 second animations, and Multi-stage Motion that can combine up to 5 motion segments. Characters can be automatically rigged, animated, and exported without remapping.

    #Product updates#Video#English text is being prepared.

  13. cnBetaProduct release
    AI score62

    1Password launches AI agent password authorization feature: agents can log into accounts but cannot read passwords

    1Password has launched a credential authorization feature for AI agents. After an agent proves its identity with a unique digital key, it sends a request to the user, who approves or rejects it on their own device. Once approved, 1Password fills the login information directly into the website login interface; the password is not handed to the agent in text form. The feature will support agents and tools such as Meta Muse, ChatGPT Work, ChatGPT Dots, Claude Duet, Grok Bot and Town, and platforms including Browserbase, Browser Use and Perplexity are also advancing integration. The original text also notes that even if an agent cannot see the password, it may still perform actions the user does not want after approval, so users need to carefully review authorization requests.

    Same story · 13 reports

    #Product updates#Agents#Safety / alignment

  14. cnBetaProduct release
    AI score79

    Anthropic launches Claude Duet agent app, focused on proactively handling professional work

    Anthropic has launched Claude Duet, an agent app powered by Claude Opus 5.5 that can continuously handle tasks in the background, running even when users leave their phone or computer. Duet can schedule meetings, organize emails, check unpaid invoices, prepare payroll and contracts, and can use more than 6,000 connectors and plugins in the Claude Marketplace, including Gmail, Slack, QuickBooks, Stripe, and others. When actions are taken on behalf of the user, Duet first shows what it plans to execute and waits for approval; it is currently launched as a research preview, offering standalone iOS and Android apps and a Claude.ai web version, and in the first phase is available only to Claude Pro and Max subscribers in the United States, who must join a waitlist.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  15. Google DeepMindFirst-partyProduct release
    AI score52

    Google DeepMind, together with WRI and IIASA, releases ForTy 2020, a 10-meter-resolution global forest type map

    Google DeepMind and Google Research, together with the World Resources Institute (WRI) and the International Institute for Applied Systems Analysis (IIASA), released Forest Typology (ForTy) 2020, which classifies every 10×10-meter parcel of land worldwide into six forest and land-use types, serving as a key baseline for the European Union's deforestation regulation. The team trained the model on about 1.7 million labeled locations and used AlphaEarth Foundations embedding vectors as a key input; in validation at more than 9,000 randomly sampled points, the forest type classification accuracy was 90%, and the accuracy in distinguishing forest from non-forest was 95%. A companion year-by-year time series from 2017 to 2025 is also openly available and can be used free of charge under an open license on Google Earth Engine.

    A close-up of a 3D voxel-style world map focusing on the blocky terrain of South America.

    #Papers/Research#English text is being prepared.#Google

  16. TechCrunch AIProduct release
    AI score80

    Anthropic Launches Claude Duet Agent App

    Anthropic has released Claude Duet, an agent app for professionals and small businesses that can generate daily briefings, create dashboards, and run tasks in the background on its own virtual machine, requesting user approval when needed. The app is based on the Claude Opus 5.5 model, can connect to more than 6,000 connectors and plugins in the Claude Marketplace, and can shop online on users' behalf via Stripe's Link digital wallet or Shopify's Shop Pay; sensitive steps such as payments require user approval, and it does not involve securities trading or transfers between accounts. Claude Duet is available in research preview for Pro and Max users on iOS, Android, and the Claude.ai web version, and requires joining a waitlist.

    Same story · 44 reports

    #Product updates#Agents#Anthropic

  17. The Verge AIProduct release
    AI score65

    1Password launches 1Password for Agents, letting AI agents log in without touching passwords

    1Password launches 1Password for Agents, allowing AI agents to log in to accounts without seeing or handling user passwords. When an agent reaches a login form, it authenticates to 1Password with a unique digital key; the app notifies the user which account credentials are needed and why, and after the user approves on their own device, 1Password fills in the login information, in a flow similar to two-factor authentication. The feature is designed to support agents such as Meta Muse, ChatGPT Work, Dots, Claude Duet, Grok Bot and Town, and partners including Browserbase, Browser Use, Perplexity and Ramp are also integrating.

    Same story · 13 reports

    #Product updates#Agents#English text is being prepared.

  18. 36Kr · AIProduct release
    AI score66

    Microsoft adds Guardian Agents to Project Perception to supervise enterprise Agents

    Microsoft has added a supervisory Agent called Guardian Agents to its security system Project Perception, used to inspect the behavior of other Agents, investigate anomalies, and tighten permissions or halt operations according to enterprise-defined rules. The plan requires first establishing an independent identity for the Agent performing the task, minimizing authorization on a per-task basis, and revoking it afterward, while also expanding Defender for Endpoint's identification scope to 60 types of locally running AI Agents. Project Perception is currently still in preview for a small number of invited customers, billed by Security Compute Unit (SCU).

    #Product updates#Agents#Safety / alignment

  19. TMTPost · AGIProduct release
    AI score65

    Microsoft adds Guardian Agents to Project Perception to supervise enterprise Agents

    Microsoft has added Guardian Agents to its security system Project Perception; they inspect the behavior of other Agents, investigate anomalies, and act according to rules preset by the enterprise. The plan requires first establishing an independent identity for the Agent performing the task, granting only the permissions needed for the current task, and revoking them once the task ends; it also expands the detection scope of Defender for Endpoint, which can discover 60 kinds of locally running AI Agents. Project Perception is still in preview for a small number of invited customers and is billed by Security Compute Unit (SCU).

    #Product updates#Agents#Safety / alignment

  20. heise onlineProduct release
    AI score59

    Perplexity Computer supports running Qwen 3.6 locally, no longer dependent on the cloud

    Perplexity has added a local model option to its macOS agent system Perplexity Computer, which can run Qwen 3.6 or its own post-trained PPLX Pro, no longer requiring dependence on cloud models and a constant internet connection. Local operation requires an Apple Silicon device with at least 36 GByte RAM recommended, and is available only to Pro, Max, or Enterprise subscribers; it also offers Hybrid Compute, which can call more than 15 cloud-based frontier models through a local orchestrator.

    #Product updates#English text is being prepared.#Agents