Anthropic Adds Shareable Dashboard Feature to Claude Code

Anthropic has added a new feature called "Artifacts" to Claude Code for paid plan users of "Claude Team" and "Claude Enterprise." This feature outputs AI work results as interactive web pages that are updated in real-time, and by sharing the URL with team members, multiple people can simultaneously monitor progress. It is characterized by the ability to visualize in-house deliverables such as dashboards and app designs on the spot.

Previously, engineers needed to set up separate infrastructure or create screenshots and explanations to share development work status with non-technical team members. Artifacts eliminates this process by having AI directly read existing codebases or connected data sources and automatically generate web pages. The generated pages update their content in real-time as the agent works, and the latest state is reflected while the URL remains unchanged.

There are also considerations for security and management. Since a new version history is recorded each time a page is updated, it is possible to roll back to previous states or track changes from both desktop and mobile devices. Engineers can continue their work on their device's terminal while team members in remote locations view the same page and understand the situation in real-time. It is worth noting that Anthropic first introduced Artifacts to its consumer-facing chatbot in the summer of 2024, and this time the feature has been integrated into the CLI (command-line interface) and desktop applications.

A similar feature is also offered by OpenAI as an update to its "Codex" platform. OpenAI's "Sites," released about two weeks ago, is a web hosting feature for enterprises and is designed as a full-fledged platform capable of continuously deploying applications on Cloudflare infrastructure. In contrast, Anthropic's Artifacts takes a philosophically different approach by adopting a "stateless" design that dynamically generates pages from the session context, rather than maintaining persistent backend infrastructure.

These moves by both companies demonstrate that competition over AI-powered enterprise work environments is expanding beyond the developer domain. In response to the challenge of how to smoothly share information between engineering teams and management/planning departments, major AI companies are presenting different solutions. Which approach better fits corporate needs depends on whether persistent infrastructure is required or whether priority is given to real-time sharing during development.