Anthropic Turns Claude Code Sessions Into Live Team Dashboards

Introduction

A coding session used to be a private thing. You typed in a terminal, the work happened, and your teammates saw the result only after you packaged it up. Anthropic has quietly shipped one of the more practically useful enterprise AI features in recent memory. It is called Claude Code Artifacts, and it takes that private session and turns it into a live webpage anyone on your team can open.

What Happened?

Anthropic launched Claude Code Artifacts in beta for Team and Enterprise subscribers. The feature converts AI coding sessions into live, interactive HTML pages that can be shared across an organization through private URLs.

The idea builds on something Anthropic already had. The original Artifacts system launched in June 2024, letting users generate static outputs like code snippets, documents, and simple visualizations inside Claude. By mid-2025, hundreds of millions of Artifacts had been created. This new version goes further. It generates live HTML pages that pull from multiple data sources and update in real time.

Here is how it works in practice. When you are working in a Claude Code session, Artifacts lets you surface that work as a custom HTML page hosted at a shareable URL. Teammates can open it and watch it update in real time as Claude Code does its work and as the connected data and code change.

Anthropic showed off a real scenario. An engineer asks Claude Code to investigate user drop-offs since a software release. The agent runs a database read, builds an interactive funnel dashboard, and finds that Pro accounts stall at the export sheet. It then proposes interface fixes, updates the charts as the code changes, and generates a secure link a manager can open on mobile.

Why It Matters

The bigger story is competition. Anthropic's update came a little over two weeks after OpenAI shipped a similar enterprise hosting feature called Sites for its Codex platform. The two companies are taking different paths, though. OpenAI is building a platform-as-a-service for durable, full-stack web apps. Anthropic is building a stateless canvas.

There is a tradeoff buyers should weigh. Both Anthropic and OpenAI chose closed, proprietary licensing for these workspaces. Neither Claude Code Artifacts nor Codex Sites can be forked or self-hosted, and enterprise clients do not own the underlying rendering engine. Your team builds on infrastructure you do not control.

There is also a quieter strategic angle. Once a team is sharing Artifacts URLs in Slack and building workflows around live AI-generated dashboards, switching costs go up. That stickiness is the real play here.

What to Watch Before You Roll It Out

If you are on a Team or Enterprise plan, this is worth testing, but go in with eyes open. Every Artifact is private to its author by default and cannot be made public to the broader internet. That answers the obvious security question, but the lock-in question stays. You cannot move these dashboards off Anthropic's infrastructure, so treat them as internal tools, not permanent systems you own.

Start small. Pick one recurring report your team rebuilds by hand every week, like an incident tracker or a funnel monitor, and let Claude Code generate it as a live Artifact. If it saves real time and your compliance team signs off on the hosting model, expand from there.