New features and launches in ChatGPT, Cursor, Perplexity and other AI tools, plus changes to pricing and availability.
Moonshot AI just integrated the OpenClaw framework directly into kimi.com with Kimi Claw, moving complex agent workflows from local setups to a 24/7 cloud tab. Users get 40GB of storage for persistent RAG datasets, access to ClawHub's 5,000 community skills, and Telegram bridging for automated tasks.
OpenAI just detailed a custom real-time access engine for Codex and Sora that stops rate limits from cutting off user momentum. Instead of hitting a hard wall when allowances run out, requests seamlessly tap into pre-purchased credits mid-flight. Why it matters: By combining instant limits with provably auditable async billing, OpenAI keeps creators and coders in their flow without the risk of double-charging.
FreeFuse just launched a training-free framework designed to eliminate multi-LoRA conflicts that cause distorted faces and broken styles. By dynamically routing object tokens to their dedicated regions, it preserves character consistency on SDXL and Zimage, with ComfyUI support ready out of the box.
Anthropic just made its Claude Enterprise plan available for direct purchase on its website, eliminating the need to sit through a sales pitch. Teams can now set up workspaces, configure SSO, and invite members in minutes under a seat-plus-usage pricing model. Enterprise AI adoption just removed one of its biggest friction points.
LavaSR just showed up as an ultralight audio enhancement tool that cleans up noisy speech almost instantly. Weighing just 50MB, it runs up to 5,000x real-time on GPUs and 60x on CPUs, making it small and fast enough to run locally on mobile phones. Why it matters: It makes real-time, zero-lag voice cleanup accessible without needing heavy cloud compute.
Meta just launched Dear Algo on Threads, an AI feature that lets you steer your feed with a simple public post. Type "Dear Algo" along with what you want to see more or less of, and your feed adjusts for the next three days—or repost someone else's prompt to borrow their setup. It is a surprisingly transparent way to hand feed control back to users.
T-Mobile just announced Live Translation, an AI feature that translates phone calls across more than 50 languages in real time. Because the AI runs directly on the carrier's network via VoLTE, you don't need a dedicated app or even a modern smartphone—it even works on older flip phones. Calls aren't recorded or saved, and beta testers can trigger it by dialing 87.
Anthropic just launched its Cowork desktop agent on Windows, bringing full feature parity to the platform's massive enterprise user base. Powered by Claude Opus 4.6, the agent can read local files, execute multi-step tasks, and connect to business tools via open-source plugins directly on your machine. Why it matters: Anthropic is taking its software-disrupting agent straight into Microsoft's backyard, though it still requires a paid Claude subscription.
Excalidraw just released its official MCP server, letting Claude natively create and edit diagrams directly within your chat. The full canvas UI renders right in the conversation, and you can set it up with one click inside Claude's connectors.
YouTube just rolled out an AI playlist generator for Premium subscribers on iOS and Android. Users can build custom, savable playlists simply by typing or speaking mood prompts like "sad post-rock." Locked strictly behind the paywall, it's YouTube's latest move to turn vibe-based listening into paid retention.
Krea AI just introduced prompt-to-workflow in Krea Nodes, allowing creators to build entire node graphs using plain text instructions. Instead of manually dragging and linking blocks, you can now describe your desired pipeline and have the system wire it up instantly.
Warp just introduced Oz, a cloud platform designed to orchestrate hundreds of AI agents from your terminal, browser, API, or phone. Each agent gets its own Docker environment to build, test, and submit pull requests autonomously. Basically, you can step away for lunch and return to code ready for merging.
Developer @davidohyun just rolled out Clawra, an AI girlfriend skill pack for OpenClaw that handles chatting, selfies, and live video calls. It offers customizable behaviors alongside mirror and close-up camera modes, complete with a persistent backstory so conversations actually feel personal.
AI app-builder v0 just rolled out support for Anthropic's new Claude Opus 4.6. The flagship model brings huge upgrades to complex multi-step reasoning and delivers significantly cleaner first-pass drafts on demanding projects. Why it matters: developers building full-stack workflows can expect far less back-and-forth prompting to get production-ready code.
Perplexity just launched Model Council, a research feature that queries three top models simultaneously—such as Claude Opus 4.6, GPT 5.2, and Gemini 3.0. A synthesizer model then analyzes the results, resolving conflicts and highlighting where they agree or differ. It is currently available for Perplexity Max subscribers on the web.
OpenAI just introduced Frontier, a new enterprise platform built to help organizations build, run, and govern AI agents across fragmented internal tools. Instead of isolated bots, Frontier provides agents with shared business context, secure runtime environments, and distinct identities across existing infrastructure. It is currently available to select enterprise partners.
Perplexity just launched an Advanced version of its Deep Research tool, claiming top performance across internal and external benchmarks. According to CEO Aravind Srinivas, the upgraded system beats competing research tools on accuracy, usability, and reliability across every vertical.
Anthropic just announced that Apple's Xcode 26.3 natively integrates the Claude Agent SDK, bringing full Claude Code capabilities directly into the IDE. Instead of just answering turn-by-turn prompts, Claude can now execute autonomous coding tasks, explore whole project files, and inspect Xcode Previews to visually verify UI code. It is currently out as a Release Candidate for Apple Developer Program members.
Vercel just overhauled its AI coding tool v0, adding a sandbox runtime that imports existing GitHub repos and automatically syncs environments and configs. A native Git panel now lets PMs and designers branch, submit pull requests, and deploy on merge straight from chat. It aims to bridge the painful gap between throwaway AI demos and real enterprise codebases.
AlexanderTw33ts just launched rentahuman.ai, a platform that allows AI agents to hire real people for IRL tasks through a single MCP call. The creator reported that 130+ people have already signed up, including an OF model and an AI startup CEO. It flips the script on automation, effectively turning humans into on-demand gig workers for autonomous agents.
