OFICIAL Cloudflare Blog

Everything we launched during Agents Week

What happened
Based on Cloudflare Blog · Aug 10, 2026

Cloudflare concluded Agents Week with a series of announcements aimed at supporting the development and deployment of AI agents, emphasizing infrastructure, security, and lifecycle management for agentic software.

Everything we launched during Agents Week
Cloudflare Blog — Cloudflare
Key points
·
This post is also available in Deutsch, Español, Español (Latinoamérica), Français, Italiano, 日本語, 한국어, 繁體中文, and 简体中文.
·
Over the last year or so, we set out to explore what this shift means for developers and customers building AI-native apps and the infrastructure needed to support them.
·
As agents become more capable and autonomous, the challenges extend beyond the models themselves — to identity, communication, orchestration, memory, observability, and security.
·
Over the past week we’ve shared how the company is bringing those pieces together across the Cloudflare platform to serve an Agentic Internet.

Cloudflare introduced a new runtime called @cloudflare/computer designed specifically for agents, enabling direct communication between Python and JavaScript Workers. The company also highlighted improvements in serving large models more efficiently while maintaining quality, reliability, and safety. Additionally, Cloudflare announced a simplified method for tracking usage and costs across its self-serve products, along with support for hosting voice AI backends or real-time voice agents using Cloudflare Workers.

The company unveiled the Agent Development Lifecycle (ADLC), which replaces the traditional Software Development Lifecycle (SDLC) to guide agents from prototype to production. Cloudflare introduced features such as live tracing and replay of agent runs, human-in-the-loop approvals for production deployments, and distributed tracing in local development environments. The announcements also included secure transaction capabilities for agents through Wallets, programmable CI/CD pipelines, and AI-powered automation to maintain code standards and streamline development workflows.

Cloudflare extended its Zero Trust framework to include agents, providing a secure method for agents to access resources and services on behalf of users. The company shared how it applies AI internally to enhance team productivity without compromising security. Cloudflare open-sourced the platform used internally for building apps, automating work, and safely accessing systems. The updates also included tools for attributing AI activity to real users and systems, helping customers monitor anomalies and spend spikes, and offering control over risky tool calls to prevent unwanted changes.

Cloudflare defined the concept of an Agentic Internet, outlining how website owners, publishers, and agents can collaborate to create an Internet that serves both humans and autonomous agents effectively. The announcements emphasized the need for infrastructure and tools that enable seamless cooperation between humans and agents in digital environments.

Original source → Deals on Clipraptor.com →