Search “Search”

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

Web Answer on Amazon Bedrock AgentCore is now generally available

Web Answer on Amazon Bedrock AgentCore is now generally available. Agents can obtain synthesized answers with source citations in a single tool call, without having to build their own search, retrieval, synthesis, and citation extraction pipeline. It returns results as an MCP tool through AgentCore Gateway; each question triggers a fresh search and page reading to verify facts, and it searches again when evidence is insufficient. Web Search still returns ranked URLs, titles, and snippets, leaving the model and synthesis process under the user's control.

NVIDIA Releases AI Factory Operations Agent to Troubleshoot DGX Cluster Failures with Multi-Agent Collaboration

NVIDIA has released the AI Factory Operations Agent, which collaborates through multiple domain sub-agents to locate root causes across the compute, network, hardware, and scheduling layers, and provides recommendations through a chat interface for its own DGX SuperPOD operations. The framework is built on the NVIDIA NemoClaw blueprint and OpenShell, is fully locally deployed, uses the local open-source NVIDIA Nemotron model for inference, and runs on DGX H100, DGX B200, and DGX GB300 production clusters. The framework includes sub-agents such as Job Recovery, Hardware, Observability, Research, Kubernetes, and Cluster Management, and integrates MCP tools such as web search, code execution, and Grafana dashboards. Developer examples and open-source code are now available.

Mastra Launches Knowledge Shared Knowledge Base for Agents

Mastra has released Knowledge, providing agents with a shared document library organized by pages that can work with signals to bring relevant knowledge into context. Knowledge controls read and write permissions across three scopes: org, resource, and thread. When tools are enabled, agents gain five tools—knowledge_search, knowledge_list, knowledge_read, knowledge_page, and knowledge_note—that can search, read, and write pages. This feature requires @mastra/core@1.82.0 or later and is currently in alpha, so the API may change between versions.

Claude launches Enterprise Managed Authentication (EMA), letting admins connect plugins org-wide via their identity provider

Claude launches Enterprise Managed Authentication (EMA), which lets admins connect plugins for the entire organization through their identity provider, so members no longer need to search for, sign in to, and authorize each plugin themselves; the capability is based on an open extension to MCP. Claude also supports automatically connecting to MCP servers that accept Claude as a trusted issuer when users sign in to Claude; this auto-connect feature is in beta for Team and Enterprise plans and already supports Amplitude, Atlassian, Datadog, Dropbox, Figma, Granola, HubSpot, Linear, Mixpanel, Notion, PagerDuty, Semrush, and Supabase. If an organization's identity provider does not support EMA, Claude can vouch for each signed-in member directly, and once an admin configures it, eligible members can connect automatically.

What Is Call Transcription? A Practical Guide for Businesses

Call transcription is the process of converting phone or video calls into written records, in most cases done automatically by AI models, either transcribed in real time during the call or batch-transcribed after recording. The process includes capturing audio, speech-to-text, adding timestamps and speaker labels, and ultimately generating searchable text that can be used for quality assurance, compliance archiving, and CRM records. Voice AI agents built with ElevenAgents save a transcript with every conversation.

MINISFORUM Showcases Full-Scenario Edge AI Solutions at the 2026 "Internet Light" Expo

At the 2026 "Internet Light" Expo, MINISFORUM presented local AI hardware solutions, including AI workstations, edge workstations, and AI NAS. The booth's focal point was a four-unit cluster composed of 4 MS-S1 MAX units; a single unit supports up to 128GB of unified memory, can allocate 96GB as VRAM, and can run models with tens of billions of parameters on an intranet. The local AI NAS N5 MAX offers up to 200TB of storage expansion and can be combined with local AI to build a knowledge base and perform semantic search.

Zyphra proposes hierarchical attention ZHA, achieving sub-quadratic complexity sparse attention

Zyphra proposes hierarchical attention ZHA, which selects relevant tokens for each query by constructing and searching a compressed tree of the KV cache, achieving sub-quadratic complexity under an appropriate top-p threshold. ZHA only requires KL distillation with a few billion tokens to adapt an existing pretrained model, adding only 81,920 trainable parameters on Qwen3.6-35B-A3B; on RULER-13 256K, the wide tree reaches 96.2% and the deep tree 93.3%, on LongBench v2 the deep tree reaches 63.4%, higher than dense attention's 58.0%, and on SWE-bench Verified it is comparable to the dense baseline. Its GPU performance model predicts it is faster than dense attention at about 48K tokens and faster than MSA at about 480K tokens, with roughly 32–35x speedup relative to dense attention at a ten-million-token context.

Lakebase Postgres Generally Available on Google Cloud

Databricks announced that its fully managed serverless Postgres database, Lakebase Postgres, is generally available on Google Cloud, supporting autoscaling, scale-to-zero, second-level branching, and point-in-time recovery. It shares the same storage layer with analytical data, eliminating the need to maintain ETL pipelines, and provides unified governance through Unity Catalog; multi-availability-zone endpoints offer a 99.99% SLA, while single availability zones offer 99.9%. Lakebase Search delivers 2x the vector and full-text search throughput of pgvector at 4x lower cost.

Tripo Launches 3D Junkyard Model Recycle Bin

Tripo introduces 3D Junkyard, created by @ReinerTao: on desktop, you can drag in models and they are placed by similarity; search with text or images and jump to matching results; models can be retrieved and reused under CC BY 4.0; Claude Code, Codex, and Cursor can also search and contribute models via MCP. The project is built on Tripo for Tripothon S1.

Hands-on Test of the Top 10 DeepSeek Harness Plugins: A Configuration Guide for Building a Personal AI Workbench

More than half a month after DeepSeek Harness was released, a large number of plugins have emerged in the community. This article hands-on tests the top 10 most popular plugins and provides installation commands. The list covers the ecosystem entry point dsh-market, the visual enhancement modlens (3.7k GitHub stars), multi-source search, long-term memory, cost monitoring, and more; dsh-web-all has the most stars at 6.4k. The author recommends that beginners first install the marketplace and visual plugins, then add interface and scenario-based tools as needed, to avoid dependency conflicts caused by installing everything at once.

Modem Launches X Integration: Turning User Tweets into Trackable Feature Requests

Modem has launched an X integration that can search X for posts about its own product, filter out spam and jokes, categorize them by topic, and push them to the #bots-social channel in Slack. Over the past 30 days, Modem read 30,000 X posts and distilled hundreds of actionable bug reports and feature requests. Feedback on the same request from X, Slack, tickets, and sales calls is merged into a single feature request, linked to a Linear issue and a GitHub PR, and once the fix ships, it lists who to notify and where to reply. The integration is included in paid subscriptions and can be enabled under Settings → Integrations → X.

Gemini Notebook to Launch Connectors, Initially Supporting Only Google Drive

Gemini Notebook is about to gain the Connectors feature, but initially only Google Drive will be available as a connector. Users can search Google Drive files directly in chat and can disable the feature as needed; officials say the connector only searches workspace documents and original files will not be modified. More Connectors are expected to be added later.

AI/ML API launches Jev Router, automatically selecting models by difficulty

AI/ML API has released Jev Router, an MCP connector for agent tools. It uses TypeSafe's Jev to assess the difficulty of each request and then automatically selects the model and tools. Simple requests are routed to lightweight models such as GPT-6 Luna, while high-difficulty reasoning requests go to Claude Opus 5.5, avoiding paying flagship-model rates for simple tasks. Users only need to provide the task; Jev Router first selects the model and tools such as web search and page extraction, then returns the result.