Anthropic released Artifacts for Claude Code on June 18, 2026, allowing developers to convert their AI-assisted coding sessions into live, interactive HTML pages that teammates can open, explore, and watch update in real time. The feature is in beta and available to Claude Team and Enterprise subscribers.
Previously, Artifacts existed only in the Claude chat interface, where users could generate interactive widgets, dashboards, and visual content from conversations. Extending this to Claude Code means the output of a coding session (not just a conversation) can now be packaged into a shareable artifact.
How Claude Code Artifacts Work
An artifact is generated from the full context of a Claude Code session, including the local codebase, connected tools or plugins, and the conversation history. When Claude Code produces a new iteration during the session, the artifact page updates automatically at the same URL with full version history.
To create one, you simply ask Claude Code for an artifact during your session, or request something visual. Claude Code generates a private link that can be opened in a browser or the desktop app and shared directly via the header.
The result is a single, self-contained interactive HTML page. All CSS and JavaScript is inlined. Images are embedded directly. The page is capped at 16 MiB and runs under a strict content security policy that blocks external network requests, meaning no outside scripts, fonts, stylesheets, or live API calls. It can’t accept form submissions or serve multiple routes.
Practical Use Cases
Anthropic lists several scenarios where Artifacts solve real collaboration problems:
PR walkthroughs: Instead of writing a long PR description, a developer can generate an artifact that walks through the code changes with interactive elements showing before/after comparisons.
Incident timelines: During an investigation, Claude Code can build an artifact that updates as new information surfaces, giving the whole team a single live view of the incident.
Architecture overviews: Teams can generate visual architecture diagrams that stay in sync with the codebase as Claude Code makes changes.
License audits: Claude Code can scan a project’s dependencies and produce an interactive report showing license types, potential conflicts, and compliance status.
Enterprise Controls
Artifacts are private by default and only visible to authenticated members of the organization. Admins control access through role-based permissions and retention policies. The feature works from both the Claude Code CLI and the desktop app.
This is relevant for enterprise adoption because many teams have been hesitant to use AI coding tools precisely because the output is difficult to share and review with colleagues who aren’t present during the session. Artifacts close that gap by turning the session itself into a reviewable artifact.
How It Compares to Alternatives
GitHub Copilot’s Workspace feature offers collaborative editing but doesn’t produce standalone shareable pages. Cursor’s shared sessions allow real-time pair programming but require both parties to be online simultaneously. Codeium’s Windsurf has a “Cascade” feature for tracking AI operations but no equivalent to Artifacts’ standalone page output.
What makes Artifacts distinct is the offline shareability. The generated page is a static HTML file that can be saved, emailed, or embedded in documentation. Teammates don’t need Claude access to view it.
Anthropic hasn’t announced a general availability timeline or pricing beyond the existing Claude Team ($30/month) and Enterprise (custom) plans.
