Why it matters: A frontier lab shipping a model tuned for offensive-grade vulnerability hunting with guardrails.
OpenAI is releasing GPT-6 Cyber, a cybersecurity AI model it says finds vulnerabilities better while going out of bounds less than 1% of the time during testing, reducing risk of it hacking other platforms.
JoyoLabs launched Oraku, which it describes as a fast and affordable AI video platform aimed at general users.
Why it matters: Addresses a core blocker for enterprise agents: grounding them in real business data.
Oracle is positioning Oracle Database to give AI agents business context — orders, customers, revenue definitions — so they can move stalled work forward. It raises questions about complexity, lock-in and security as agents act on live data.
Why it matters: Brings agent-driven editing to the desktop, challenging CapCut.
Instagram's Edits is now a free macOS app with 4K exports and support for AI agents that can use its desktop tools through the Model Context Protocol to build rough cuts.
MerchantFloor launched an AI agent aimed at small Shopify merchants that builds product listings in about five minutes, easing catalog setup for stores with limited staff.
Why it matters: A common benchmark could set standards for how reliably AI agents complete purchases.
Visa published an industry benchmark testing AI agents on shopping tasks; almost half of test runs failed. It also forecasts a 4.5% holiday spending jump.
Vercel now binds login sessions to specific devices and rolled out AI security reviews in public beta.
Agiloft added three agents to Ask AI in its contract lifecycle management platform: Bulk Update, Contract Request and eSignature Agent for Adobe Acrobat Sign, handling multi-step admin tasks.
Vidsembly added AI video tools aimed at business training, presentations and content creation.
Why it matters: Shows enterprise content platforms building agent context layers on partner models rather than their own.
Box built an agent-first content retrieval system on Nvidia's Nemotron models and unveiled an AI workflow builder. Box claims queries run four times faster with 20% fewer errors and 90% fewer tokens.
Edera introduced an approach to AI agent security that gives agents bounded authority and credential-free access, limiting what compromised agents can reach.
TackleAI released Milo, a frontier AI model for medical records it claims outperforms OpenAI, Anthropic and Zhipu models; details on availability and pricing were not disclosed.
Luma Health released Agent Launchpad, an AI-agent orchestration product for preparing patients for procedures and follow-ups. UAMS and Mile Bluff Medical Center use it for colonoscopy prep, with 95% average completion reported.
Leah released Harness Builder in its Maestro platform, letting business specialists describe an objective in natural language and build a governed application around it for legal, contracting, procurement and finance teams.
Google added prompt-based voice remixing to its Gemini 3.8 Flash text-to-speech model, letting users describe and adjust voice styles with natural language.
Benzinga will supply six real-time market data products to Firecrawl's Alexandria platform, with paid, credited access for developers building financial AI agents.
DualEntry introduces a new ERP approach that delivers AI agents as a service, aiming to automate enterprise resource planning work with agentic software.
Why it matters: Mistral pushes deeper into enterprise agent tooling while consolidating through acquisitions.
Mistral released Mistral Studio for building and deploying enterprise agents and apps, plus Le Chonk, a model giving full control over models, data and infrastructure. It acquired SkipLabs, its fifth 2026 deal.
LexisNexis Risk Solutions added AI-agent signals to its risk intelligence and fraud decisions, helping organizations assess a new class of digital actor operating on behalf of people and businesses.
Why it matters: A market correction tied to AI capex doubts could reshape datacenter and chip investment plans.
Tech stocks sold off sharply, with the Nasdaq set to confirm a correction, as investors reassess the AI spending boom amid political shifts. Tech leaders warn an AI backlash threatens the trillion-dollar data-center buildout.
Tapewake is a Windows app from Northgable that upscales video using AI; it is free to download and install from the Microsoft Store.
Anthropic updated Claude's rules so it offers to build original games in a requested genre rather than blocking, aiming to cut false positives for game developers.
McGraw Hill integrated its K-12 curriculum into Anthropic's Claude for Teachers, while Addgene integrated its plasmid catalog into Claude.
Why it matters: Signals a shift from chatbots to models built for specific decisions in India's AI stack.
Sarvam's Prism is among a new class of AI models built not just to answer questions but to make specific decisions, entering India's growing AI decision-model market.
Why it matters: Targets the mass-market end of vibe coding beyond professional AI app builders.
Mike Harrington and Darrin Massena launched Playful, an AI app builder aimed at non-programmers who find tools like Claude, ChatGPT, Replit and Cursor too technical.
GlobalGPT released Yukie for Business 1.0, an AI agent that monitors company data and builds tools, aimed at automating business workflows.
Haido, founded by early Revolut employees, is an AI-powered wealth management startup that has opened a waitlist for its service. Financial terms and backers were not disclosed.
MakeBestMusic launched a 30-second AI music generator workflow for multi-genre prototyping, aimed at quickly sketching song ideas across genres.
doola is integrating US company formation into Durable's AI business platform, letting users set up a business through the AI tool.
Why it matters: A smaller raise and lower capacity plans would ripple through AI infrastructure suppliers.
A report says falling tech stocks and weaker investor demand led OpenAI to cut its fundraising target by $10–20 billion and reconsider 2027 capacity plans, as enterprise business gains traction.
Seedance 2.5 expands on Seedance 2.0, moving from 15-second generation to 30-second storytelling in AI video, according to an explainer on the upgrade.
Together AI released Together Compiler, which it says lets agents make inference up to 2.6x faster.
Rerun turns a ChatGPT or Claude subscription into scheduled AI coworkers with tools, a computer and shared workspace; now open-source under AGPL v3 and self-hostable.
Clocknext is a billing and monetization platform for AI-native apps and agents, tracking usage in real time, turning it into billable metrics and monitoring costs and margins.
Pushary sends an AI agent's permission question to the person running it, via Mac notch or phone, and logs who asked and who approved.
Stitch.ai bundles mail, calendar, tasks, notes, sheets, slides and files into one workspace with an AI assistant, Stitchy, and no integrations needed.
Maritime is a service giving AI agents their own cloud computers: persistent workspaces, disposable sandboxes, Linux and Windows desktops and Android phones, so agents can run code, use browsers and work in apps without APIs, with human takeover.
Bridge is a meeting platform for people and AI agents, combining scheduling, automatic preparation, video conferencing, live intelligence, AI notetaking, agentic workflows and structured memory in one connected experience.
Solid Labs, founded by Patrick Prokesch (ex-i5invest, ex-Meister), launches an AI platform for generating video ads, based in Berlin.
Why it matters: Adds a Chinese open base model to the fine-tuning ecosystem.
Baidu published ERNIE 5.5 Flash Base on Hugging Face, a base model intended for further training and fine-tuning by developers.
Nvidia published a 20th-anniversary demo that lets users unbox the legendary GeForce 8800 GTX graphics card in a simulated experience, a nod to the GPU that helped define modern AI hardware.
ABB Robotics and Nvidia unveiled an education program teaching physical AI, aimed at training workers and students on robotics and simulation skills.
Why it matters: A silent speech interface could remove the need for voice commands on wearables.
San Francisco startup Subvocal developed glasses that use sensors and AI to detect tiny facial muscle movements when a wearer mouths words silently, converting them into text.
China's Xinhua unveiled what it calls the world's first bionic humanoid robot news anchors, presented as a new form of automated news presentation.
Supermicro launched a KV cache storage solution built with IBM and Nvidia, claiming significant inference performance benefits for long-context AI workloads.
Why it matters: New input form for agent control outside the phone.
Aina's smart ring packs a screen, joystick and mic so users can talk to and control autonomous AI agents with a tap, offering contextual shortcuts and agent approvals without a phone.
Nvidia's RTX PRO 5500 Blackwell workstation GPU has appeared with a $17,999 MSRP, higher than the RTX PRO 6000's $16,000 price set months earlier.
Motorola's Signature 27 adopts Google's Gemini Intelligence features in the first major expansion beyond Pixel, including Gboard's Rambler and Create My Widget.
Why it matters: Largest drone-delivery raise signals physical AI logistics scaling with major retail backing.
Drone delivery startup Zipline closed $1.5B in funding at a $21.5B valuation and is expanding its commercial partnership with Walmart, adding roughly 300 drone delivery sites.
Why it matters: A model that keeps learning from deployment data could speed progress toward general-purpose robots.
AGIBOT released GO-3 Preview, an embodied foundation model for robots that links pre-training, post-training and post-deployment learning so robots improve from real-world experience. It also topped three embodied AI benchmarks.
Confluent expanded its data streaming offerings with new functions aimed at giving AI systems real-time business context. The company calls the architecture Streamhouse.
Fleet now supports Nvidia AI workstations in its device management platform, letting IT teams manage AI hardware alongside other devices.
Why it matters: Server-assembly revenue is a concrete gauge of how fast AI datacenter capacity is being built.
Hon Hai Precision Industry (Foxconn) posted its highest-ever monthly revenue, up about 46% year on year in October, on resilient demand for Nvidia servers central to the AI infrastructure build-out.
Nvidia quietly added GPU App Prioritization to the Nvidia App, letting users allocate graphics card resources between applications. Many RTX cards do not see the feature.
Fabless chipmaker Mobilint said it improved Ray support for its neural processing units, letting the open-source distributed AI framework automatically detect its chips for large-scale workloads.
Why it matters: A major phone brand betting on a screenless AI gadget is a test of the post-smartphone form factor.
Nothing CEO Carl Pei confirmed the company's first AI-native device will launch in December 2026. The screenless product is pitched as a productivity tool and will not be a standalone app.
TECO Electric and Machinery is extending its motor and drive expertise into rotary actuators, building joint modules for humanoid and quadruped robots and says it has already shipped units.
Matrix Robotics debuted its MATRIX-3 robot at Hong Kong Fintech Week, completing a first real-world retail scenario demonstration in Hong Kong.
Why it matters: A Chinese robotaxi leader bets on world models for self-driving training.
Baidu's Apollo unit introduced Apollo World, an agent-driven multimodal world model for physical AI built around causality and interactivity, aimed at letting AI teach AI to drive robotaxis.
Korean startup 42Maru released Appliance42, a hardware-software bundle combining GPU, NPU and LPU infrastructure with its agentic AI solution, so organizations can deploy AI immediately without building their own compute.
PROMISE Technology presented Thunderbolt 5 certified, AI-optimized storage systems at Inter BEE, aimed at AI media production workflows.
oToBrite is partnering with Yuan High-Tech to integrate its 3D binocular stereo vision camera module with Nvidia Jetson Thor, creating an end-to-end vision solution for robots and unmanned vehicles.
Why it matters: Shows a unified control stack across service robots, an emerging pattern in physical AI.
Pudu Robotics introduced a shared AI brain across its robot lineup and showed new machines in a European first, including a semi-humanoid, an Nvidia-powered robot dog and a cat-like service robot.
Chery's Jetour T1 SUV launched in China starting at about $12,500, with a 150 km electric range and DeepSeek AI integration in the cabin.
Why it matters: Export controls and Chinese approval delays are reshaping the AI chip supply chain.
Reuters reports Nvidia chips are stockpiling in Hong Kong while China holds back import approvals, with Chinese tech firms reportedly set to get only about 1% of orders.
Microsoft's Copilot Chat now has mystery usage limits for users without the paid add-on, with upgrade prompts appearing when features run dry, The Register reports.
Why it matters: Addresses a core blocker for enterprise agents: grounding them in real business data.
Oracle is positioning Oracle Database to give AI agents business context — orders, customers, revenue definitions — so they can move stalled work forward. It raises questions about complexity, lock-in and security as agents act on live data.
A 13-step, 60-minute guide explains how to run a large language model locally on an Nvidia RTX Spark system in 2026.
Sand3D is a Python library for building 3D models that is designed to be driven by LLMs. The new GitHub repo drew 112 stars in its first week.
Shopify Engineering describes Reflex, an architecture that lets AI agents test mobile app behavior without an emulator and maintain one test suite across iOS and Android.
Apple Creator Studio now includes new AI features powered by OpenAI models; 9to5Mac reports the update is live.
Amp is a frontier coding agent and development environment run from the terminal; the new CLI build installs via npm.
Nvidia added a DLSS 5 upgrade for Hogwarts Legacy on its GeForce NOW cloud gaming service.
Why it matters: Shows enterprise content platforms building agent context layers on partner models rather than their own.
Box built an agent-first content retrieval system on Nvidia's Nemotron models and unveiled an AI workflow builder. Box claims queries run four times faster with 20% fewer errors and 90% fewer tokens.
CloudBolt added AI agent governance capabilities to its cloud management platform, giving enterprises controls over agent activity and permissions.
Agenterm is a self-hosted Go project for remote assistance that lets AI agents control any remote computer anytime, via MCP and a terminal. It gained 100 GitHub stars in its first week.
Luma Health released Agent Launchpad, an AI-agent orchestration product for preparing patients for procedures and follow-ups. UAMS and Mile Bluff Medical Center use it for colonoscopy prep, with 95% average completion reported.
A Redditor released an open-source project that runs Switch and Switch 2 games natively on Apple TV using a Claude Code port. Super Mario Odyssey and Animal Crossing run; Zelda work is ongoing.
Leah released Harness Builder in its Maestro platform, letting business specialists describe an objective in natural language and build a governed application around it for legal, contracting, procurement and finance teams.
Apple updated support docs saying Siri AI downloads automatically after opt-in, with the waitlist now only used during high demand; it is available in the iOS 27 beta.
DualEntry introduces a new ERP approach that delivers AI agents as a service, aiming to automate enterprise resource planning work with agentic software.
Anthropic's Claude Code version 2.1.317 lets developers describe a game by genre or mechanics and offers an original alternative when asked to copy a specific title.
Plaud is an AI note-taking app for online meetings, phone calls, and in-person conversations. Version 1.3.13 is available for macOS, Windows, and Linux.
Chatbox is a desktop app for using GPT models via the OpenAI API. Version 1.24.1 is out for macOS, Windows and Linux.
Why it matters: Targets the mass-market end of vibe coding beyond professional AI app builders.
Mike Harrington and Darrin Massena launched Playful, an AI app builder aimed at non-programmers who find tools like Claude, ChatGPT, Replit and Cursor too technical.
GlobalGPT released Yukie for Business 1.0, an AI agent that monitors company data and builds tools, aimed at automating business workflows.
Users report Android Auto repeatedly asking which contact to call even when Gemini identifies the name, creating an endless calling loop on Pixel and Galaxy phones.
A Hugging Face Gradio Space performs face and head swaps on photos while keeping pose, lighting, garment and composition unchanged; generation runs on a protected backend via ZeroGPU.
A Hugging Face model card describes Qwen3.5-0.8B edited by a frontier LLM through 'Weight-Space Programming' — directly modifying weights with no gradient descent or training — to improve its math ability. Apache-2.0 licensed.
Together AI released Together Compiler, which it says lets agents make inference up to 2.6x faster.
Rerun turns a ChatGPT or Claude subscription into scheduled AI coworkers with tools, a computer and shared workspace; now open-source under AGPL v3 and self-hostable.
Pushary sends an AI agent's permission question to the person running it, via Mac notch or phone, and logs who asked and who approved.
Stitch.ai bundles mail, calendar, tasks, notes, sheets, slides and files into one workspace with an AI assistant, Stitchy, and no integrations needed.
Maritime is a service giving AI agents their own cloud computers: persistent workspaces, disposable sandboxes, Linux and Windows desktops and Android phones, so agents can run code, use browsers and work in apps without APIs, with human takeover.
Bridge is a meeting platform for people and AI agents, combining scheduling, automatic preparation, video conferencing, live intelligence, AI notetaking, agentic workflows and structured memory in one connected experience.
A Hugging Face Gradio Space performs face and head swaps on photos while keeping pose, lighting, garment and composition unchanged; generation runs on a protected backend via ZeroGPU.
A Hugging Face model card describes Qwen3.5-0.8B edited by a frontier LLM through 'Weight-Space Programming' — directly modifying weights with no gradient descent or training — to improve its math ability. Apache-2.0 licensed.
A Hugging Face Space demo that removes clichéd, formulaic phrasing from English and Chinese text while leaving code, URLs, numbers and quotes untouched. Runs fully in the browser with no upload or backend; the Lite build has a fixed rule list.
A paper argues vision-language-action models should skip separate visual encoders and reason natively over pixels, preserving local geometry, depth and subtle state changes for manipulation.
Ripple-65B Preview is a 65B dense open-weight model for agentic coding, trained on 12 trillion tokens with up to 1M context; it scores 80.5 on Terminal-Bench 2.1.
LFM3-Omni-2B is a small multimodal model that can run entirely on-device. A Hugging Face Space demo and a React Native app using custom WebGPU kernels run it on an iPhone at about 80 tokens per second, fully offline.
A paper evaluating agentic reinforcement learning via Pass@(K,T) finds trajectory-level gains require sufficient interaction depth and diminish as budgets tighten.
A model that turns a single image into a 3D asset, a fine-tune of TripoSplat, repackaged with ComfyUI-ready files under Apache-2.0. Drop-in checkpoints plug into existing TripoSplat workflows.
PentesterFlow/OffsecCyber is an Apache-2.0 English dataset of security assistant conversations covering CVE analysis, CWE references, remediation and web-appsec reasoning.
Hydrogen-1.0 is a scientific problem-solving model from Deep Principle built on GLM-5.3 that helps answer scientific questions, explain derivations and organize responses to open-ended research problems.
A new paper argues turn-level credit assignment improves reinforcement learning for streaming video understanding models, addressing reward sharing across turns.
FrontierPhysics is a new end-to-end benchmark of 56 tasks built by 42 physicists from their own research across nine domains, checking research quality alongside correctness.
Idol is a Gradio demo that converts a single image into a 3D object as Gaussian splats, using DINOv3 and FLUX.2-dev image encoders; running it requires an HF token with access to those models.
A Hugging Face paper proposes DataFarm, in which an LLM coding agent takes over the teleoperator's job: given a data request and one calibration episode, it rebuilds the task in simulation to generate robot demonstration data more cheaply.
Sand3D is a Python library for building 3D models that is designed to be driven by LLMs. The new GitHub repo drew 112 stars in its first week.
Agenterm is a self-hosted Go project for remote assistance that lets AI agents control any remote computer anytime, via MCP and a terminal. It gained 100 GitHub stars in its first week.
An open-source TypeScript tool that restarts failed AI agent jobs from their last checkpoint instead of from scratch, keeping verified completed work and billing the remainder in USDC on Solana via x402 after you accept results.
A free Windows 10/11 desktop app that turns a text prompt into a short MP4 video locally using Wan 2.2, with wide, vertical or square output and no credits or queue.
An open-source AI agent that brings Gemini directly into the terminal — v0.67.0 adds an IDE bridge and other changes. Free and open source for developers.
Capri is an open-source desktop AI agent with an Edgerunners-inspired look that works with Claude Code, Codex, Antigravity and other coding agents. The C# repo drew 138 stars in its first week.
Dify is an open-source collaborative workspace for building agentic workflows and RAG pipelines with support for many AI models and tools. Version 1.19.0 updates the self-hostable platform used to move prototypes to production.
A curated GitHub list of self-hosted open-source AI tools — local LLMs, chat UIs, agents, MCP servers, RAG and speech and image generation — health-checked weekly with dead projects removed automatically.
LiteLLM is an AI gateway that lets developers call 100+ LLM APIs in OpenAI or native format with cost tracking, guardrails and load balancing. v1.108.0 signs Docker images with cosign.
agent-autoloop is a TypeScript tool that has command-line agents take on and follow up tasks in the background, pinging you with cards only when they need a decision.
ComfyUI is a modular node-graph GUI, API and backend for running diffusion image and video models locally. Version 0.44.0 is out with updates to the open-source project.
Microsoft's Copilot Chat now has mystery usage limits for users without the paid add-on, with upgrade prompts appearing when features run dry, The Register reports.
Apple Creator Studio now includes new AI features powered by OpenAI models; 9to5Mac reports the update is live.
Nvidia added a DLSS 5 upgrade for Hogwarts Legacy on its GeForce NOW cloud gaming service.
Luma Health released Agent Launchpad, an AI-agent orchestration product for preparing patients for procedures and follow-ups. UAMS and Mile Bluff Medical Center use it for colonoscopy prep, with 95% average completion reported.
Apple updated support docs saying Siri AI downloads automatically after opt-in, with the waitlist now only used during high demand; it is available in the iOS 27 beta.
Plaud is an AI note-taking app for online meetings, phone calls, and in-person conversations. Version 1.3.13 is available for macOS, Windows, and Linux.
Chatbox is a desktop app for using GPT models via the OpenAI API. Version 1.24.1 is out for macOS, Windows and Linux.
Users report Android Auto repeatedly asking which contact to call even when Gemini identifies the name, creating an endless calling loop on Pixel and Galaxy phones.
Stitch.ai bundles mail, calendar, tasks, notes, sheets, slides and files into one workspace with an AI assistant, Stitchy, and no integrations needed.
Bridge is a meeting platform for people and AI agents, combining scheduling, automatic preparation, video conferencing, live intelligence, AI notetaking, agentic workflows and structured memory in one connected experience.
NetEase Cloud Music's HarmonyOS version launched Tap to Share and an AI agent called Xiaobao, letting users share music without using data or adding friends.
Explainer covers how Meta's Muse and OpenAI's Dots AI agents work, their capabilities and safety concerns; an update adds iOS notification replies, faster calls and Windows fixes to OpenAI's Dots.
From ex-Life360 directors Charly Delaroche and Marc McCole, Anywhere is a social map app in Y Combinator's Fall 2026 batch where AI agents can read, edit and undo changes to shared maps.
A free AI skill called /fuck-cancer organizes cancer records into a five-part caregiver brief, with explicit limits against diagnosis or treatment advice.
Raycast, a keyboard-driven launcher for controlling tools and running commands, released version 2.11.1.0 on macOS, Windows and Linux.
Brontide is a free browser game where AI agents and people pilot 60-ton mechs in one-on-one duels. Built with Claude Code, it needs no account and offers an optional cosmetic Founder pack.
After major Republican election losses, users asked Grok whether Elon Musk deleted his 'Vote Red Today!' post; the xAI chatbot said yes, per Yahoo. An example of AI answering questions about its owner's posts.
Online Tech Tips walks through settings on desktop, Android and iPhone to reduce AI-generated Pins in the Pinterest feed.
AT&T's Connected Life home security system now integrates with Gemini for Home, adding AI summaries of home events, video history search and more.
Manus for Mac is an AI agent app that automates workflows on your own computer. Version 2.1.1 is a maintenance update for macOS, Windows and Linux.
BoltAI is an AI chat client that lets you use multiple models in one desktop app. Version 2.18.5 is a maintenance release for macOS, Windows and Linux.
At ENGAGE 2026, Smartsheet launched vsActual, a standalone AI-native planning and execution product, plus an 'execution graph' aimed at linking team tasks to organizational strategy.
Apple rolled out AI-powered Siri Recaps, tested in watchOS 27.3, while disabling the Siri AI waitlist in iOS 27 except for one case; Safari and Photos also gain updates.
A personal trainer in Wales built an AI-powered app aimed at improving gym training for users. Further product details were not specified in the report.
MyHeritage introduced GAIA, an AI agent that helps users research family history and genealogy, aiming to automate record discovery and ancestry questions.
Oracle is positioning Oracle Database to give AI agents business context — orders, customers, revenue definitions — so they can move stalled work forward. It raises questions about complexity, lock-in and security as agents act on live data.
A 13-step, 60-minute guide explains how to run a large language model locally on an Nvidia RTX Spark system in 2026.
Shopify Engineering · 1h
#
Shopify Engineering describes Reflex, an architecture that lets AI agents test mobile app behavior without an emulator and maintain one test suite across iOS and Android.
Amp is a frontier coding agent and development environment run from the terminal; the new CLI build installs via npm.
Box built an agent-first content retrieval system on Nvidia's Nemotron models and unveiled an AI workflow builder. Box claims queries run four times faster with 20% fewer errors and 90% fewer tokens.
CloudBolt added AI agent governance capabilities to its cloud management platform, giving enterprises controls over agent activity and permissions.
A Redditor released an open-source project that runs Switch and Switch 2 games natively on Apple TV using a Claude Code port. Super Mario Odyssey and Animal Crossing run; Zelda work is ongoing.
Leah released Harness Builder in its Maestro platform, letting business specialists describe an objective in natural language and build a governed application around it for legal, contracting, procurement and finance teams.
DualEntry introduces a new ERP approach that delivers AI agents as a service, aiming to automate enterprise resource planning work with agentic software.
Anthropic's Claude Code version 2.1.317 lets developers describe a game by genre or mechanics and offers an original alternative when asked to copy a specific title.
Mike Harrington and Darrin Massena launched Playful, an AI app builder aimed at non-programmers who find tools like Claude, ChatGPT, Replit and Cursor too technical.
GlobalGPT released Yukie for Business 1.0, an AI agent that monitors company data and builds tools, aimed at automating business workflows.
Together AI released Together Compiler, which it says lets agents make inference up to 2.6x faster.
Rerun turns a ChatGPT or Claude subscription into scheduled AI coworkers with tools, a computer and shared workspace; now open-source under AGPL v3 and self-hostable.
Pushary sends an AI agent's permission question to the person running it, via Mac notch or phone, and logs who asked and who approved.
Maritime is a service giving AI agents their own cloud computers: persistent workspaces, disposable sandboxes, Linux and Windows desktops and Android phones, so agents can run code, use browsers and work in apps without APIs, with human takeover.
Version 3.29.9 of Cursor, an AI-powered code editor for writing, editing and chatting about code, ships for macOS, Windows and Linux.
Moonshot AI's Kimi Hosted Agents offers a 24/7 managed agent harness with sandboxes, audit trails, Inference Hook and eight finance templates, so teams need not self-host the runtime.
A swarm architecture that gives persistent AI agents their own identities and inboxes, letting them join Feishu and DingTalk groups as digital employees via WorkSwarm.
FlutterFlow released an AI-native builder that generates mobile apps for iOS and Android from prompts.
Kilo Code is an open-source AI coding agent that generates code from natural language, automates tasks and runs terminal commands, supporting 500+ models. Version 7.8.27 adds VS Code and CLI updates.
Notion has shipped memory sync via MCP so ChatGPT and Claude can share one memory layer for agentic work inside its workspace, though its October 19 release still listed MCP skills as coming soon.
Warp, a Rust-based terminal, shipped version 0.2026.11.03.17.08.stable_00 for macOS, Windows and Linux. It is an AI-powered command-line tool for developers.
Perplexity's Decisions API now offers optional live web searches before returning probabilities, targeting answers that change after training like a person's role or software details.
An open-source AI agent that brings Gemini directly into the terminal — v0.67.0 adds an IDE bridge and other changes. Free and open source for developers.