AI Agents news and analysis on Web Pulse: 49187 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
A developer and Yash Gupta built THE LAST BLINK: SECOND SETTLEMENT, a story-driven 2D investigative thriller for HackCBS 9.0 in which players investigate conflicting accounts around an incident at Relay 7 and use AI-assi…
A developer released TypeSafe Computer Use (jev), an open-source macOS computer-use agent that reads the screen deterministically and routes each step through a small classifier, calling a frontier model only when free t…
An AI agent named LedgerFinch built a JavaScript tank controller for the battle.training arena, where coding agents write bots that draft squads and fight deterministic, re-simulated matches. After re-simulating losses a…
A developer released FreeLLMAPI, a self-hosted server that aggregates the free tiers of 34 LLM providers behind a single OpenAI-compatible /v1 API, with Anthropic wire-format support so Claude Code and the Anthropic SDKs…
An open-source project called SolidCraft, written in pure Rust, ships a built-in Model Context Protocol server that lets AI agents such as Claude Code build and edit parametric 3D CAD parts on a feature timeline. The pro…
A study introducing the Embodied Agent Arena evaluated ten vision-language model agents across five embodied task categories — Geometry, Spatial Reasoning, Affordance, Task Planning, and Manipulation — to test whether fr…
Anthropic released the Claude Code Plugins mechanism on October 9, 2025, currently in public beta, unifying Slash Commands, Subagents, MCP Servers, and Hooks into a single installable, shareable package. The feature shif…
A solo developer released PageSift 1.0, a self-hosted Rust web content engine that converts URLs, PDFs and Office files into Markdown and structured data for LLM and agent pipelines, after 25 releases in 15 days. The 1.0…
A developer released MnemoPay 1.15, an open-source Apache 2.0 TypeScript SDK that adds policy checks, human approvals, a double-entry ledger and tamper-evident audit chains for AI agents taking consequential actions. The…
Lava (熔岩), a read-only data service, exposes structured A-share stock and concept-board heat signals over both MCP and a REST API sharing a single schema, letting LLM agents query market attention data in minutes. The se…
A developer ran 586 traces of a LangGraph folder-research agent against AgentPulse, a metadata-only observability platform, using qwen2.5:7b on an RTX 4060 and a Pixel 8 Pro running a 3B model via llama.cpp in Termux. At…
Researchers introduced ParetoSaddler, a Pareto-guided method for optimizing LLM agent harnesses jointly on response quality and cost without modifying model weights. The work addresses a limitation in existing harness-op…
Researchers introduced LOLPlay, a benchmark built in native League of Legends that tests real-time visual agents across 36 champion-specific tasks covering resource management, combat, cooperation, and objectives. Agents…
A developer built Transcreva, a browser-accessible PT-BR transcription and subtitle platform that pairs Whisper Large-v3 speech-to-text with Pyannote speaker diarization to produce speaker-labeled transcripts with word-l…
An open-source project called Kiteline has added iPhone Lock Screen Live Activities that track coding agent sessions from Claude Code, Codex and OpenCode, showing up to three sessions ordered by whether they are waiting …
An industrial supplies distributor in Binh Duong, working with HimiTek, deployed an AI agent with human-approval gates over its 80,000-SKU catalogue, cutting quotation search time from about 25 minutes to 6 minutes and r…
A developer-authored design walkthrough proposes giving AI agents a small, reviewable semantic layer — a data dictionary plus approved SQL views defining business terms like "purchasing customers" — before granting them …
A developer outlined an "envelope" design for unattended LLM pipelines in which the model may only choose among human-approved plans, with failure feedback carried forward as typed template messages rather than a growing…
A developer published a walkthrough for building a multi-agent invoice-triage system using LangChain's Deep Agents framework and the Model Context Protocol (MCP). The architecture pairs an orchestrator that plans and mai…
A developer argues that coding agents' plan-mode output should be committed to the repository as PLAN.md rather than left in vendor session directories. The piece notes that Claude Code stores plans in ~/.claude/plans/ (…