AI Article Tags

Browse articles by professionally maintained tags. Every tag opens a focused collection with its own editorial introduction and closely related published articles.

Flowith An AI creation and knowledge workspace built around agent workflows: FlowithOS turns browsing, retrieval and service calls into one execution chain, while iOS 2.0 hands control to a proactive voice agent. Kagi Unpack Kagi's paid search model: how Lenses, the Universal Summarizer and the Assistant chain into one research workflow, and what the ad-free, privacy-first design changes for daily use. SAM Audio Walk through Meta's SAM Audio approach to audio separation: how text, visual and time-range prompts stack to pin down a target sound, and what the PE-AV encoder underneath contributes. Bloom Explain how Bloom auto-generates evaluation scenarios around a single behaviour: the four-stage scoring pipeline, the seed mechanism for reproducibility, and how it divides work with Petri. GLM-4.7 Focus on GLM-4.7's open weights and agentic coding: how finer thinking control and multi-step tool use steady long-horizon tasks, and what its default role in the Coding Plan means for adopters. Hunyuan3D Break down Tencent Hunyuan's 3D model family: how layered 3D-DiT sculpting reaches 1536³ geometry, what P3-SAM segmentation and X-Part generation add, and where 3D Studio's eight-view reconstruction fits. Unitree See how Unitree splits the work between robot dogs and humanoids, and what the 4D LiDAR and open interfaces on the Go2 mean for research labs, classrooms and industrial inspection. CodeRabbit Explain how CodeRabbit generates line-level suggestions and applicable patches during pull requests, and what changes once MCP integration brings business and operations context into the review. Qwen3-ASR Sort out the trade-offs behind Qwen3-ASR in multilingual recognition, noise robustness and long audio, and what the ForcedAligner component and the CLI toolkit each add to the chain. HunyuanImage A look at the open HunyuanImage 3.0 and its Instruct variant: how an 80B MoE text-to-image model follows long prompts, and where image-to-image editing and multi-photo fusion still have limits. Google Antigravity Break down how Google Antigravity puts long-running agents rather than autocomplete at the centre of coding, and what remote session control changes for teams supervising multi-hour refactors. Baidu Records Baidu's two lines from model to product: Huixiang video generation on MuseSteamer and GenFlow 2.0's parallel agent orchestration, with the Wenxin family underneath. Opera One Lays out Opera One's mix of the Aria assistant, tab islands and a free VPN, and what split screen, naming and colour solve when many tabs are open, without piling on extensions. MyPrompt Breaks down MyPrompt's prompt library, its categories and entry structure, and how creators take ready-made prompts by theme to ease weak keywords and unstable output across several models. OpenAI Prism Looks at OpenAI Prism, the cloud LaTeX workspace that folds drafting, compile preview and multi-person collaboration into one project, and what its editor assistant takes over in paper writing. OpenAI Agents SDK Traces OpenAI Agents SDK from v0.10.5 to v0.11.1, showing how the agent loop, tools and handoffs carry multi-step tasks and where the framework sits in a developer's stack. Cowork Breaks down Cowork, Anthropic's folder agent, showing how it brings Claude Code's reach to tasks like tidying files or turning receipts into tables, and where its platform and privacy limits lie. DeepSeek-OCR Explains how DeepSeek-OCR compresses context with visual tokens, breaks down its resolution options and Markdown conversion, and shows which structured document tasks it fits. Ray-Ban Meta Traces the Ray-Ban Meta line from the classic frame to new prescription models, covering the conditions for photos, open-ear audio and Meta AI, and the trade-off between looks and ecosystem reliance. Verdent Lays out Verdent's desktop and VS Code forms, showing how parallel sub-agents chain planning, execution and verification into a task-level flow, and where its control stands in large-repo refactoring. Mixboard Looks at Mixboard, Google Labs' AI concept board, and how Gemini plus image models turn templates, prompts and your own pictures into a moodboard you can show, with multi-board projects and wider file support. KAT-Coder Shows how KAT-Coder chains requirement breakdown, defect localisation and unit-test generation through agentic coding, and what it trades off between speed and quality in large repositories. Rokid Breaks down Rokid's split between display-equipped Rokid Glasses and the big-screen AR Lite, showing how display, battery and comfort are traded off inside a light frame. Adobe Photoshop Tracks how Photoshop reaches chats and browsers through its ChatGPT app and a Chrome extension, and how far AI retouching actually goes under subscription and regional limits.