AI Tools news and analysis on Web Pulse: 37021 curated articles tracking the latest AI Tools developments, tools, and research, updated continuously from vetted sources.
Google released an early-access Google Home Model Context Protocol (MCP) server that lets AI-powered development tools and virtual assistants inspect home structures, monitor device states, execute control actions, and a…
Jared Palmer, vice president of engineering at Cognition, released a Qwen3.5-based generation of the open-source Kev decision-model project on September 20th, comprising 0.8B, 4B and 9B parameter variants that return typ…
A new non-profit site, pressaudit.org, is using Alibaba's Qwen 3-32b large language model to process thousands of news articles and analyze them by region, public interest, and partisan lean, according to the project's l…
OpenAI released its Agents API into public beta on September 10, exposing the same harness that runs Codex — built-in context compaction, session persistence, and multi-step tool coordination — at no extra fee beyond sta…
A developer used Hermes, a homelab agent, and Claude Code to build a 101-track, 6.2-hour Spotify playlist for a child's baptism party, played hands-off via djay Pro on an iPad. The first attempt failed because the Spotif…
AI-generated code accounted for 17.25% of all Linux Kernel patches in September, according to The Lunduke Journal, which reported the figure on X. The publication also said 1,634 code submissions written by AI were made …
A developer who has reviewed recordings from live AI voice agent deployments argues that callers hang up not because of voice quality or speed but because agents are scoped to intents they cannot finish and cannot hand o…
Avestura released Boxy, an MIT-licensed zero-radius design system and agent skill that enforces square-corner interfaces through 15 lint rules and a `boxy-check` CLI that parses CSS, HTML, JSX, Vue and Svelte and exits n…
A developer reviewed the source code of more than 100 open-source projects that use TypeSafe's JEV decision model, cataloguing the results in the awesome-jev repository. The review found that the most valuable patterns w…
CreditWatch 1.0.0, a local-first macOS menu bar utility for monitoring prepaid cloud credits, was released on 21 September 2026 with Vast.ai as its first and only supported provider. The app stores a Vast.ai key in Keych…
Hume published its Voice Replication Leaderboard on September 10, 2026, testing 11 voice cloning models on 25 reference voices with 7 prompts, where Fish Audio s2-pro led the vendors covered at 4.03 on a 1-to-5 speaker-s…
Developer Sylwia Lask presented WebMCP, an experimental browser API that lets websites explicitly expose tools for AI agents to call, at AGNTCon + MCPCon Europe before an audience of over 2,000 attendees. She demoed AI C…
Maxim AI's Bifrost, an open-source AI gateway written in Go, is being positioned as a centralized MCP gateway for Anthropic's Claude Code CLI, aggregating multiple Model Context Protocol tool servers behind a single auth…
Jared Palmer released Kev, a family of small decision models built on Qwen3.5 in 0.8B, 4B, and 9B sizes, with pretrained weights, training code, and evaluation data available on GitHub. Kev's API matches TypeSafe's Syste…
A developer built an omnichannel shopping assistant that shares long-term memory between a text chat agent and a real-time voice agent using Amazon Bedrock AgentCore Memory, Strands, and Amplify Gen 2. The system uses Ag…
A developer built Spellbook of Skill, an MIT-licensed collection of development tutorials and guides organized by category and difficulty, as a companion to the earlier Spellbook of Prompt. The project was created end to…
A developer reported that a local model, Ternary Bonsai 2, generated the parse_prompt function returning ParsedPromptInfo from a 180-line Markdown specification in about 30 minutes, producing 300-400 lines of Rust with 3…
A community member is seeking guidance on dataset sources and training architecture for a project that empties indoor spaces and removes specific objects from images using provided object masks. The poster asks whether p…
A medical writer built the MDR Evidence Gap Agent, a demo that queries a synthetic Class III device dossier stored in Sanity to surface claims lacking linked clinical evidence and the EU MDR requirements they affect. The…
A benchmark of 50 product-classification tasks found that replacing an agentic LLM pipeline with TypeSafe's System One model, Jev, cut mean latency 86% from 9.62 seconds to 1.38 seconds and reduced model calls 56% from 7…