Cloudflare has introduced two new features to support the development and management of AI agents: Cloudflare Agents and local tracing for Workers. Cloudflare Agents is a new platform designed for deploying and managing hosted agents, initially focusing on observability features. This platform provides agent tracing, offering direct visibility into agent behavior, including model calls, tool execution, and token usage.
In parallel, Cloudflare's development tools, `wrangler dev` and `vite dev`, now automatically capture OpenTelemetry traces for local Worker invocations. This functionality is activated when the tooling detects an agent session, directing the agent to the Local Explorer API. This local debugging API allows agents to query these traces without requiring manual SDK installation, tracing enablement, or agent configuration.
The integration of local traces provides coding agents with structured feedback directly from the development environment before code deployment. This allows agents to identify and resolve issues, such as a `POST /api/orders` returning a 500 error, by analyzing the traces. The Local Explorer API is automatically discovered by agents when they start `wrangler dev` or `vite dev`.
Agent tracing, a core component of Cloudflare Agents, measures and presents every model call, tool execution, and token used by an agent. This provides insights into an agent's actions and costs. The feature supports OpenTelemetry-compatible agent harnesses, including Think, Flue, and AI SDK.
These new capabilities build on Cloudflare's existing developer platform investments, such as the introduction of Miniflare and making local mode the default in Wrangler 3. The company views agents as another application type that benefits from its established infrastructure for model access, durable runtime, orchestration, sandboxed execution, and persistent storage.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's briefNew every morning, and the back catalogue is archived by date.
Cloudflare launched agent tracing, the initial component of Cloudflare Agents, to provide a centralized view of deployed agent sessions and detailed spans for AI agent activities within Workers. This feature addresses the challenge of debugging AI agents where application telemetry alone does not reveal internal agent behavior, offering insights into model calls, tool execution, and subagent interactions.
Cloudflare's `wrangler dev` and `vite dev` now automatically capture OpenTelemetry traces for local Worker invocations, enabling AI agents to debug issues using a new Local Explorer API. This feature provides structured feedback for agents during local development, reducing the need for manual logging and token consumption.
Cloudflare introduced Cloudflare Agents, a new platform for deploying and managing hosted agents, starting with observability features. This platform provides agent tracing to offer direct visibility into agent behavior, including model calls, tool execution, and token usage.