On August 19, 2026, Cursor expanded its cloud agents and harness so agents can run as an always-on system: they subscribe to events such as pull requests and Slack threads, hold a…
Microsoft released ThinkingBox, an open-source, MCP-compatible sandbox and paired benchmark (ThinkingBox-Bench) that evaluates AI agents by the database state they actually change…
On August 19, 2026, TrueFoundry launched TrueForge, an open-source (MIT) vendor-neutral AI agent harness on GitHub and PyPI, positioned as a cheaper alternative to Claude Managed…
On August 19, 2026, Vercel launched Vercel for Slack in public beta, an integration that lets teams collaborate on coding, investigate production issues, and run approved remediat…
On August 19, 2026, Vercel added an Instagram adapter to its Chat SDK, letting developers build AI chat agents that operate in Instagram Direct Messages with support for media, qu…
On August 18, 2026, AWS moved Amazon Bedrock AgentCore Payments to general availability, letting AI agents autonomously discover, access, and pay for paid APIs, MCP servers, and c…
On August 18, 2026, KTX launched its Skills Kit in Seoul, letting AI agents running in Claude, ChatGPT, Codex, and Cursor tap KTX market data, AI-generated intelligence, portfolio…
Pinecone announced general availability of Nexus, a knowledge engine that compiles an enterprise's proprietary documents, databases, and workflows into a governed, agent-ready kno…
On August 18, 2026, Vercel made Cline available in the AI SDK harness layer via the @ai-sdk/harness-cline adapter, built with the Cline team, letting developers drive the Cline co…
On August 18, 2026, Vercel added Z.ai's GLM-5.3 to its AI Gateway, callable with a single API key with no platform fee, automatic fallbacks, spend tracking, and per-request traces.
In mid-to-late August 2026, Google's Agent2Agent (A2A) interoperability protocol became a hosted project of the Linux Foundation-directed Agentic AI Foundation (AAIF), placing it…
On August 17, 2026, Microsoft released Agent Lightning v1.0, an MIT-licensed, open-source reinforcement-learning framework that trains and optimizes AI agents without changing the…
On August 17, 2026 Synchrony announced an enterprise collaboration with OpenAI to bring financing, rewards, and loyalty into AI-native shopping. A Synchrony ChatGPT plugin is live…
Vercel is discounting OpenAI's GPT-5.6 Sol by 50% on all requests routed through the Vercel AI Gateway from August 17 to September 18, 2026, announced in its August 18 changelog.
On August 19, 2026, Stripe officially announced an agreement to acquire OpenRouter, the AI model-routing gateway that spans 400+ models from more than 80 providers, confirming rep…
On August 14, 2026 Cloudflare added DeepSeek V4 Flash and DeepSeek V4 Pro to Workers AI, the first models on the platform with a full 1,048,576-token context window. Both support…
On August 13, 2026, Cursor introduced builds for Cloud Agents: ready-to-use copies of a project's development environment (repos cloned, dependencies installed, install scripts ru…
OpenAI previewed Ultrafast, a new API service tier that runs GPT-5.6 Sol at up to roughly 750 output tokens per second, up to 14x faster than standard processing, using Cerebras w…
On August 13, 2026, Perplexity launched its Agent API, a single programmable endpoint for web search, URL fetching, code execution, MCP connections, and finance and people search,…
On August 13, 2026, Vercel made the v0 API generally available, letting developers invoke its AI app-building agent programmatically from their own products, scripts, CI pipelines…
Writer released Palmyra X6, a flagship enterprise model priced at $2 per million input and $8 per million output tokens, alongside upgrades to the Writer Agent harness. Writer say…
On August 12, 2026, DeepSeek moved its flagship V4-Pro (V4-Pro-0813) to general availability across web, mobile, and API. The 1.6-trillion-parameter mixture-of-experts model (abou…
On August 12, 2026 Liquid AI released LFM2.5-VL-3B, a 3.1-billion-parameter open-weight vision-language model built to run on phones, laptops, and single GPUs rather than in a dat…
On August 12, 2026, Microsoft AI moved MAI-Thinking-1, its first from-scratch reasoning model trained without distilling from OpenAI or other providers, into public preview in the…
On August 12, 2026 OpenRouter launched Auto, a new automatic model router informed by the model choices of millions of users across its marketplace. OpenRouter says Auto outperfor…
Oracle Cloud Infrastructure added Background Mode to its Responses API, letting long-running AI tasks run asynchronously without keeping an active connection open. Oracle position…
On August 11, 2026 xAI (SpaceXAI) and Cursor shipped their first joint product after their merger: Grok Bot, an agent swarm in which each bot gets a dedicated cloud computer so it…
On August 11, 2026 NVIDIA released Nemotron 3.5 Lightning, an open-weight 30B mixture-of-experts model with 3B active parameters and a 1M-token context, distilled from Nemotron 3…
On August 10, 2026 Meta Superintelligence Labs released Muse Glimmer under the Apache 2.0 license on Hugging Face, a 30-billion-parameter dense multimodal model engineered for off…
OpenAI is retiring its standalone Atlas browser on August 9, 2026, nine months after launch, and moving browser-based agentic capabilities into ChatGPT and Codex. ChatGPT gains a…
In early August 2026, LangChain launched Managed Deep Agents, a service that bundles its Deep Agents harness with managed infrastructure to run production-grade, long-running agen…
On August 7, 2026 Cloudflare announced Kitesurf, a cloud-hosted browser engine written in Rust and compiled to WebAssembly that runs entirely inside Cloudflare Workers V8 isolates…
On August 7, 2026 Cloudflare introduced Radar Researcher, an AI tool that answers plain-language questions about global Internet trends and traffic with interactive charts and exp…
Cloudflare merged Workers AI and AI Gateway into one AI control plane with a shared binding and REST endpoint. A default gateway is now created automatically for observability, AI…
On August 7, 2026 Vercel made its AI Gateway and Sandbox available inside Hermes Agent, giving the agent inference across 200+ models with no token markup and isolated microVM exe…
On August 7, 2026, Vercel introduced Skill Packs, which bundle multiple agent skills into a single shareable, unlisted URL built from community skills, local folders/zips, or publ…
On August 6, 2026 AWS made AgentCore Runtime Instances generally available, letting teams run Bedrock agents on their own Amazon EC2 instances without managing infrastructure. It…
On August 6, 2026, AWS introduced Runtime instances for Amazon Bedrock AgentCore: persistent, managed EC2 infrastructure for production AI agents with multi-agent collaboration, G…
Cloudflare shipped developer-experience improvements to AI Search, its managed retrieval service that turns your files and websites into an agent-ready search engine, and previewe…
On August 6, 2026 Cloudflare introduced Kitesurf, a browser engine built for AI agents that runs entirely on Cloudflare Workers in V8 isolates. It implements the Chrome DevTools P…
On August 6, 2026 Cloudflare launched a developer preview of WebMCP, letting site owners expose tools to visiting AI agents through a single dashboard toggle, so agents can call w…
On August 6, 2026 Cloudflare launched a developer preview of WebMCP, which lets site owners make any website usable by browser-based AI agents with a single switch — no new APIs a…
On August 6, 2026 Cloudflare detailed the next generation of the Model Context Protocol: the 2026-07-28 spec, interaction model, and SDKs were rewritten to be fully stateless, so…
On August 6, 2026, GitHub released Agent Plugins 1.0, an open standard that bundles agent skills and Model Context Protocol servers into a single installable plugin that works acr…
Vercel's Chat SDK now ships a chat/workflow module where a single requestApproval call pauses a running Workflow SDK workflow and posts an Approve/Deny card until a person decides…
A cross-vendor standard called Agent Plugins 1.0, initiated by Vercel with a steering committee of Amazon, Anysphere (Cursor), Microsoft, and OpenAI plus GitHub, defines how agent…
Cloudflare open-sourced Cloudflare OS, a version of its internal AI platform that any organization can self-deploy to run AI agents and apps grounded in its own systems and knowle…
Vercel Sandbox now supports up to 10,000 concurrent sandboxes on Pro and Enterprise plans, up from 2,000, with per-minute vCPU allocation scaling automatically to 5,000. A dynamic…
Cloudflare added agent tracing to its Agents SDK, letting developers inspect production agent behavior turn by turn. Traces surface each agent turn alongside model calls, tool run…
Cloudflare introduced Cloudflare Wallets and cloudflare.pay, giving AI agents a stable digital identity and the ability to pay in stablecoins within owner-defined limits. Each age…
Cloudflare announced Cloudflare Wallets, a programmable wallet that lets AI agents make autonomous payments and carry verifiable identity on the web using the open x402 protocol,…
During Agents Week, Cloudflare announced Cloudflare Wallets and cloudflare.pay on August 4, 2026, letting AI agents identify themselves and pay for APIs, MCP tools, data, and infe…
During Agents Week on August 4, 2026, Cloudflare announced Cloudflare Wallets, which let AI agents hold stablecoins and pay for the resources they consume, and cloudflare.pay, an…
On August 4, 2026 GitHub and Microsoft released the GitHub Copilot agentic harness as stable for both .NET and Python inside the Microsoft Agent Framework, letting developers use…
Vercel introduced skill packs on skills.sh, letting developers bundle multiple agent skills into curated, shareable collections distributed as a single URL. Packs can be assembled…
Vercel shipped agent-browser, a new @agent-browser/eve extension that lets eve agents navigate the web like a human. Agents get tools for navigation, reading page content, clickin…
Cursor released Google Workspace plugins that give its coding agents direct access to Gmail, Google Drive, Calendar, Docs, and Sheets. Agents can pull context, draft and update fi…
Cloudflare added a Playground to the Browser Run dashboard that lets developers try Quick Actions against live browsers without deploying code first. From the Playground they can…
On July 31, 2026, Cloudflare released a provisioning API for its Media over QUIC (MoQ) network, letting developers create their own isolated relays with control over publisher and…
Vercel MCP now supports the 2026-07-28 Model Context Protocol specification, giving newer clients a stateless request model and hardened authorization while 2025-protocol clients…