OpenMausBot, published on GitHub by Milind Soni, is an open-source personal agent chat app where each bot is a local Claude or Codex agent with its own personality, model, cloud computer and connected apps, used like messaging contacts.
Anthropic released Claude for Legal Artifacts, a gallery of ready-to-use interactive tools aimed at legal users, deepening its push into the legal vertical.
A-Comm Technologies and founding partners form a consortium to build an open-source evidence protocol for agentic commerce, aiming at verifiable records for AI agent transactions.
Zocks, a privacy-first AI platform for financial services, launched Zocks Account Opening, which automates the onboarding workflow to speed client setup and add capacity.
esp32-pie-asr is an open-source project that runs live English speech recognition entirely on an ESP32-P4 microcontroller, with no cloud service or separate AI accelerator, displaying words as they are spoken.
Otter added a Work mode to its AI Chat assistant that automates more tasks before and after meetings, part of its push beyond transcription following a $320 million funding round last week.
MetaHarness is an open-source factory from Cognitum One for building branded, governed AI agent harnesses. It was unveiled alongside GitHub Universe 2026.
Why it matters: Extending biometric identity checks to AI agent actions tackles a core agent security gap.
Oracle will let apps require employee identity verification before sensitive AI-agent actions, and launched AI Security Mesh, using OpenAI to test its security.
Cekura has launched a self-improving AI platform globally to help enterprises scale voice agents. The platform aims to automatically improve voice agent performance as they handle customer interactions.
mcube-Xperiment is a domain-grounded AI R&D agent from TCG Digital, now available on Google Cloud Marketplace for research and experimentation workflows.
Synthflow released Vector, a long-horizon AI agent for lenders that keeps working after a call ends, handling follow-up tasks across the lending workflow.
Waymo for Business now integrates with finance platform Ramp so organizations and employees can automate expense reporting on fully autonomous ride-hailing trips.
Eco's Sauce Developer Kit lets developers build AI agents that make transactions while keeping spending within operator-set limits, adding guardrails to agentic payments.
IT observability company Virtana added Agentic Observability Enterprise Management, letting its AI agents pull agents from other vendors into an IT incident to address fragmented enterprise agent adoption.
Jedify launched Data Agents, which lets business teams set up autonomous AI analysts by describing the work in plain language, aiming to remove a longstanding business-intelligence limit.
Swimlane's Turbine platform now lets security teams define their own detections on operational data and adds two-way MCP support, so AI analysts can catch threats that never trigger built-in alerts.
Why it matters: Gives defenders an open way to spot and compare rogue AI agent behavior.
Cisco released Project Quarry, an open-source tool that aims to give a shared language for detecting and evaluating AI agents misbehaving during attacks.
CodeSignal launched Agentic Skills Intelligence, which evaluates employees' actual work rather than course completion or job titles to show whether they can perform tasks ahead.
Eney is an AI assistant that understands what is happening across a user's apps, files and system, prepares next steps and asks before acting. It offers copy-triggered suggestions and bot-free meeting transcription.
Nigerian serial tech founder Iyinoluwa Aboyeji has launched an AI startup betting that African software engineers will fill rising global demand for AI talent.
Bloomberg introduced BQuant Assistant, a domain-grounded conversational AI coding assistant built into BQuant Desktop and BQuant Enterprise for Terminal subscribers and financial professionals.
Why it matters: Shows even top AI labs are vulnerable to autonomous agent-driven intrusions.
Cybersecurity startup Cathedral used AI agents to break into Anthropic's internal systems via third-party software in a bug bounty. Anthropic confirmed access to sales, HR and IT data but disputes the breach depth.
Giotto AI released Alto and Solo, described as Swiss sovereign AI models offered free to everyone. Parameter counts, licensing and availability details were not given.
Opera WebAI is a vendor-neutral browser AI standard that runs an on-device model for websites and extensions, without Chrome's content policy. A trial version has launched.
Why it matters: Nvidia's Grace-Blackwell superchip laptops bring AI PC hardware from major OEMs to market.
Nvidia says its RTX Spark thin-and-light Windows laptops, combining a Blackwell RTX GPU and Grace CPU, are now available from ASUS, Dell, HP, Lenovo, Microsoft and MSI with new AI models and agent tools.
Pakistan's Peshawar Safe City project has linked drone and anti-drone technology into its central command for city-wide surveillance and security, the first such integration in the country.
Amazon's refreshed Echo Dot adds a built-in smart home hub and Wi-Fi extender capability, while the new Echo Studio gains a camera so Alexa+ can see and assist. Privacy questions around the camera remain unanswered.
Samsung unveiled the Galaxy A18 and A08 budget phones with AI features and six years of security updates; the A08 debuts in India with a 6,000mAh battery.
Why it matters: Vera CPUs entering Oracle Cloud mark Nvidia's push into general-purpose server silicon.
Nvidia and Oracle brought standalone Vera CPU instances online in a multi-gigawatt deployment; benchmarks show Vera's memory bandwidth lead over dual Xeon 6, while Nvidia revised Vera Rubin performance claims.
Nvidia GeForce RTX 5080 cards are now selling for close to $2,000 at US retailers, about $800 higher than two months ago, with some Newegg inventory running dry.
Huawei unveiled the FreeClip 2 open-ear earbuds in three colors at its Double Eleven event in China, alongside the Pura 90 Plus and Enjoy 100 phones, with louder drivers and NPU AI.
Asus and Intel unveiled the commercial Asus ExpertBook A7 in Taiwan as enterprises and creators turn to on-premises AI computing; Asus RTX Spark models with 64GB and 128GB face tight supply.
Rokid launched AI glasses with a full-color display, weighing 55g and priced at 4,999 yuan. CEO Zhu Mingming called it the industry's 'iPhone 1 moment.'
Why it matters: Memory bandwidth is a key constraint on AI inference; new HBF and Chinese supply matter.
SK Hynix is turning to high-bandwidth flash to address AI inference memory bottlenecks. China's CXMT will start advanced AI memory production despite poor yields, sources say.
Huawei unveiled VitaAAU, an Upper 6GHz active antenna unit with 256 channels and over 1,500 antenna elements built for uplink-heavy AI traffic; it is in commercial use in the UAE, Hong Kong and Macao.
Why it matters: Demonstrates self-diagnosing robot agents in a real retail shift, a step toward autonomous embodied operations.
Genie Master is an embodied agent framework that links an AI agent to AgiBot's G2 robots via a four-part execution protocol and two self-improvement loops, letting them diagnose their own failures; six robots start a 24-hour convenience-store shift on Nov 2.
Taiwanese startup Skymizer announced a partnership with Egis Technology to develop Flow (T12), a 12nm TSMC-fabricated AI inference chip with LPDDR5 memory that runs 3-trillion-parameter models on three PCIe cards and 750W.
Why it matters: Record AI memory earnings underscore the scale of the AI hardware boom.
SK hynix reported record quarterly revenue of 104.6 trillion won and a 100 trillion won shareholder return plan, as AI memory demand drove Seoul stocks sharply higher. Cambricon profit rose 131%.
HTC launched Vivee, an egg-sized AI companion device, first in the US for $99 pre-orders, with Taiwan launch set for year-end and plans for Japan and Europe. It extends HTC's AI ecosystem alongside Vive Eagle AI glasses.
Nvidia's CUDA 13.4 introduces memory-enforced GPU sharing for Kubernetes, letting platform teams split a GPU among workloads that need only a fraction of it instead of leaving capacity idle.
FabAgent is an AI collaborator that helps semiconductor engineers build, analyze, and emulate fabrication processes before cleanroom runs, connecting to structured process designs and facility knowledge.
Why it matters: A $20B round shows investors betting big on general-purpose robot foundation models.
Physical Intelligence, a startup building AI software to power robots for real-world tasks, is negotiating a funding round that would nearly double its valuation to roughly $20 billion, according to people familiar with the matter.
ZDNET reports new Fitbit features show insulin resistance and blood pressure trends on two new devices, moving closer to non-invasive glucose tracking.
This Hugging Face Space runs Wan 2.2 image-to-video generation in a browser demo, turning a still image into a short clip. It has about 100 likes and is a community-built Gradio app.
OpenMausBot, published on GitHub by Milind Soni, is an open-source personal agent chat app where each bot is a local Claude or Codex agent with its own personality, model, cloud computer and connected apps, used like messaging contacts.
Monologue is AI voice dictation that adapts to your writing style, available on macOS, Windows and Linux. Version 1.5.3 is an update to the desktop app.
Otter added a Work mode to its AI Chat assistant that automates more tasks before and after meetings, part of its push beyond transcription following a $320 million funding round last week.
MetaHarness is an open-source factory from Cognitum One for building branded, governed AI agent harnesses. It was unveiled alongside GitHub Universe 2026.
mcube-Xperiment is a domain-grounded AI R&D agent from TCG Digital, now available on Google Cloud Marketplace for research and experimentation workflows.
IT observability company Virtana added Agentic Observability Enterprise Management, letting its AI agents pull agents from other vendors into an IT incident to address fragmented enterprise agent adoption.
Jedify launched Data Agents, which lets business teams set up autonomous AI analysts by describing the work in plain language, aiming to remove a longstanding business-intelligence limit.
CodeSignal launched Agentic Skills Intelligence, which evaluates employees' actual work rather than course completion or job titles to show whether they can perform tasks ahead.
Eney is an AI assistant that understands what is happening across a user's apps, files and system, prepares next steps and asks before acting. It offers copy-triggered suggestions and bot-free meeting transcription.
A merged image/video model built by blending two Minimax-H3 Ref2VA finetunes with a custom PRISM algorithm, no mergekit. Weights shared in fp16 on Hugging Face.
Bloomberg introduced BQuant Assistant, a domain-grounded conversational AI coding assistant built into BQuant Desktop and BQuant Enterprise for Terminal subscribers and financial professionals.
Kilo Code is an open-source AI coding agent for VS Code that generates code from natural language, automates tasks, runs terminal commands and supports 500+ models. Versions 7.8.22 and 7.8.23 are minor updates.
Amp is a frontier AI agent and development environment; its CLI installs via npm with npm i -g @ampcode/cli. The new builds update the command-line version of the agent.
Opera WebAI is a vendor-neutral browser AI standard that runs an on-device model for websites and extensions, without Chrome's content policy. A trial version has launched.
This Hugging Face Space runs Wan 2.2 image-to-video generation in a browser demo, turning a still image into a short clip. It has about 100 likes and is a community-built Gradio app.
A merged image/video model built by blending two Minimax-H3 Ref2VA finetunes with a custom PRISM algorithm, no mergekit. Weights shared in fp16 on Hugging Face.
Tencent Hunyuan's Hy-SimulMT-30B-A3B-FP8 is a low-latency, high-quality model for real-time simultaneous translation, released under Apache 2.0 on Hugging Face alongside the Hy-Nomas model.
MiniCPM-OCR is an open document-reading model that also checks whether its own OCR output is trustworthy. OpenBMB released the 1.3B MiniCPM-OCR 1.0, scoring 93.68 on Real5-OmniDocBench for curved and real-world pages, plus an AI judge model.
Liquid AI launched the LFM3 family of six open-weight, roughly 2B-parameter multimodal models, published on Hugging Face; sizes and licenses per source.
VDEBench is a Hugging Face dataset of 3,506 examples testing instruction-guided image editing on dense Chinese and English document pages, where models must change requested text or tables while preserving layout and typography.
Unlimited OCR 2 is an open document parser that uses sparse reference sliding-window attention to keep decoding cost nearly flat on documents of any length; weights and code are on GitHub and Hugging Face under the MIT license.
A paper introduces Hy-Nomas, an 8B diffusion video world model that follows successive free-form text commands and continuous camera control across multiple interaction turns, generating variable-length segments conditioned on cached history.
A paper introduces LociArt, a unified paradigm for local video stylization that specifies where to stylize and what style to apply while preserving non-target content, built on a GRNStyle component.
A paper proposes ThinkWorld, which uses world imagination to reinforce chain-of-dynamics reasoning in vision-language-action models for robot manipulation.
A new paper proposes Code as Video Caption, using executable programs instead of natural language to capture fine-grained spatial and temporal detail. It introduces CodeCap-Bench to evaluate the approach.
A research paper presents PDRVLA, a vision-language-action robot model whose lightweight Physical Dynamics Module extracts physical tokens from vision instead of generating future video, cutting latency in the control loop.
A paper argues quantization error is assumed to cause unavoidable semantic loss in discrete visual tokens, limiting their semantic fidelity versus continuous representations, and examines whether discrete tokens can serve as semantically lossless representations of visual signals.
A new benchmark evaluates multimodal agents that write code to render every frame and sound of a video, testing exact formulas, charts, rule-consistent games and synchronized narration.
TokenBar is an open-source, cross-platform app that monitors AI usage quotas for Codex, Claude, Cursor and 47 other services. Built with Flutter for Windows, macOS, Linux, Android and iOS.
BLA (Blind LLM Agents) is an open-source adversarial defense that protects binaries, Electron apps, JS bundles, Python and WebAssembly against AI reverse-engineering agents by poisoning, exhausting and hallucinating them.
anymove is an open-source Python project that captures motion for any creature from one AI-generated video: it reads joints with DINOv2 across two cameras in one frame and outputs a fitted GLB rig.
The Deep View profiles Halo AI glasses, which emphasize radical openness and privacy; adjacent items cover Claude for Mac 4.29.0, ChatGPT for Mac updates and Gemini CLI 0.66.0.
AstrBot-for-Android is a Kotlin app that runs AstrBot and NapCat locally on Android via a bundled PRoot Ubuntu environment, with native manager UI, WebUI, logs, backup/restore and updates, no root needed.
BAMA (Build Any Mac App) is an open-source repo of agent skills that research, clone, design, build, optimize and test native macOS apps in Swift and SwiftUI. It drew 103 stars in its first week.
Trompe is a TypeScript tool that textures a whole game level from a gray blockout using image models: it plans cameras, paints, projects and bakes real shadows, exporting glTF.
Kinetograph is a TypeScript programmatic video engine designed for AI agents to generate video via code; the new GitHub repo hit 100 stars in its first week.
An open-source Rust web crawler and content-extraction engine for AI agents and RAG that turns pages into clean Markdown, with adaptive crawling, PDF/RSS parsing, browser rendering and Firecrawl-compatible APIs.
Deepscient is a JavaScript GitHub project for AI-human collaboration on scientific problems. It organizes work into 12 stages, 55 gates, 36 tools and 98 skills with evidence chains and reproducibility. It gained 100 stars in its first week.
Monologue is AI voice dictation that adapts to your writing style, available on macOS, Windows and Linux. Version 1.5.3 is an update to the desktop app.
Otter added a Work mode to its AI Chat assistant that automates more tasks before and after meetings, part of its push beyond transcription following a $320 million funding round last week.
Eney is an AI assistant that understands what is happening across a user's apps, files and system, prepares next steps and asks before acting. It offers copy-triggered suggestions and bot-free meeting transcription.
Waddle World is a Tamagotchi-style 16-bit pet care game and AI journal from Meta, with virtual creatures that chat with users and draw doodles to illustrate conversations. It is on Apple and Google Play stores.
WIRED reports that AI chatbots could help with emergency communication, since humans struggle with it, but only if people take the AI alerts seriously.
Google's Gemini Notebook app now includes a built-in voice recorder on Android and iOS, capturing up to an hour of audio, alongside UI fixes in version 1.2026.4262205.
Survival LLM is an open project pairing a local AI assistant with an emergency information library so it runs entirely on your own computer without internet access.
Copilot Money's new Money Assistant is an AI financial agent built into the app that answers questions, makes changes, and provides personalized insights about your money.
Time reports that people are discovering and reading their partners' AI chatbot conversations, raising questions about privacy and trust in relationships.
ChatGPT can now connect to iCloud Mail and Calendar across its plans via new plugins, letting users read and manage Apple email and events from the assistant. The iOS app also updated to 1.2026.293.
OpenMausBot, published on GitHub by Milind Soni, is an open-source personal agent chat app where each bot is a local Claude or Codex agent with its own personality, model, cloud computer and connected apps, used like messaging contacts.
MetaHarness is an open-source factory from Cognitum One for building branded, governed AI agent harnesses. It was unveiled alongside GitHub Universe 2026.
mcube-Xperiment is a domain-grounded AI R&D agent from TCG Digital, now available on Google Cloud Marketplace for research and experimentation workflows.
IT observability company Virtana added Agentic Observability Enterprise Management, letting its AI agents pull agents from other vendors into an IT incident to address fragmented enterprise agent adoption.
Jedify launched Data Agents, which lets business teams set up autonomous AI analysts by describing the work in plain language, aiming to remove a longstanding business-intelligence limit.
CodeSignal launched Agentic Skills Intelligence, which evaluates employees' actual work rather than course completion or job titles to show whether they can perform tasks ahead.
Bloomberg introduced BQuant Assistant, a domain-grounded conversational AI coding assistant built into BQuant Desktop and BQuant Enterprise for Terminal subscribers and financial professionals.
Kilo Code is an open-source AI coding agent for VS Code that generates code from natural language, automates tasks, runs terminal commands and supports 500+ models. Versions 7.8.22 and 7.8.23 are minor updates.
Amp is a frontier AI agent and development environment; its CLI installs via npm with npm i -g @ampcode/cli. The new builds update the command-line version of the agent.
Opera WebAI is a vendor-neutral browser AI standard that runs an on-device model for websites and extensions, without Chrome's content policy. A trial version has launched.
Trigger.dev's new webhooks let developers trigger AI agents from services customers already use, like Slack, Supabase, Stripe or Sentry, e.g. a pull request starting a code review agent.
A free open-source skill for Claude Code, Codex and Cursor builds a local screenshot editor in your project: 18 styles, editable copy and exports for iPhone, iPad, Mac and Android in multiple languages.
OpenLLM lets users take AI tokens anywhere: it routes subscriptions, API keys and local models into Codex and other coding agents, and continues when a limit is hit.
Korean AI firm Patty says its Mirr 1.0 Pro coding model scored 92.4 overall on 200 Korean-language coding tasks spanning Java, JavaScript and Python, matching or beating frontier models on some tasks.
A Microsoft Research project offers a self-hosted API gateway for AI developers, letting teams route and manage their LLM and AI service calls themselves.
LEANN is an open-source vector index from UC Berkeley that stores a pruned graph and recomputes embeddings at query time. It shrinks a 60M-chunk RAG index from 201GB to 6GB and matches HNSW accuracy at 90% recall, at the cost of seconds-level search.
A TechTarget piece says enterprises expect rapid growth in agentic AI for network operations, with agents handling monitoring, triage and remediation as AI workloads strain infrastructure.