Claude Code Artifacts: Anthropic adds live pages (beta)
Artifacts turn Claude Code sessions into interactive, updateable web pages built from the full session context and shared inside.
TL;DR
- 01Artifacts turn Claude Code sessions into interactive, updateable web pages built from the full session context and shared inside.
- 02Anthropic announced on Jun 18, 2026 that Claude Code now supports Artifacts, a feature that converts full coding sessions into interactive web pages teams can share.
- 03The pages are generated from the session context — code, connected tools, and chat history — and update automatically at the same URL, with version history preserved.
Anthropic announced on Jun 18, 2026 that Claude Code now supports Artifacts, a feature that converts full coding sessions into interactive web pages teams can share. The pages are generated from the session context — code, connected tools, and chat history — and update automatically at the same URL, with version history preserved.
What are Artifacts in Claude Code?
Artifacts are interactive web pages created from the entire Claude Code session context, and they reflect changes automatically at a single URL. The pages include code, connected tools and chat history; when something in the session changes the page updates and retains a version history so teams can track edits and roll back if needed.
The feature mirrors the Artifacts capability already present in Claude chat and is intended for use cases such as PR walkthroughs, incident timelines, license audits and architecture overviews. Artifacts are private by default and visible only to authenticated members of an organization. Administrators manage access using roles and retention policies.
How do teams create and share Artifacts?
Teams create an Artifact by asking for one during a Claude Code session, or by requesting any visual during the session, and Claude Code returns a link that opens in a browser or the desktop app. That same link can be shared from the page header so team members can open the live view and follow updates.
Anthropic says Artifacts work with both the Claude Code CLI and the desktop app. The company made the feature available as a beta for Claude Team and Enterprise customers. The workflow is meant to capture the full session state, including connected tools, so an Artifact represents more than just a snapshot of files: it is an export of an interactive session.
Why it matters
Artifacts reduce the friction of turning an in-session conversation and tool output into a shareable, persistent artifact. Teams no longer need to stitch together screenshots, pasted logs or separate documentation to show how code, tools and chat combined to produce a result. The automated updates and built-in version history keep the shared page current and auditable, which matters for incident reviews, compliance tasks and cross-team handoffs.
Administrative controls and default privacy settings limit accidental exposure inside organizations, making Artifacts a collaboration-first feature rather than a public publishing tool. That focus aligns the feature with internal engineering workflows where context and access control are essential.
What to watch
Watch whether Anthropic expands Artifacts beyond the Team and Enterprise beta, and whether third-party integrations appear that let external systems consume or embed Artifact pages. A public roadmap or changes to retention and role controls would signal how Anthropic plans to position Artifacts across broader customer tiers.
Anthropic provided the feature details and availability on Jun 18, 2026, describing its intended enterprise use cases and the platforms it supports.
Written by The Brieftide · Source: The Decoder
The Brieftide Daily · 06:00
Briefs like this one, in your inbox every morning.
Continue reading
More in Coding AgentsAdobe creative agents arrive in Photoshop, Premiere, and more
Firefly-powered AI assistants automate multi-step production tasks across Creative Cloud and plug into ChatGPT, Claude.
CODA-BENCH benchmark: testing code agents on data tasks
CODA-BENCH places agents in a Kaggle-based Linux sandbox with 1,009 tasks across 31 communities and an average of 980 files per task.
SWE-Explore: benchmark shows AI coding agents miss key lines
SWE-Explore isolates code search from repair and finds agents hit the right files but cover only 14–19% of the lines that matter.
OpenAI acquires Ona to add persistent agents to Codex
The deal brings Ona's cloud development environments into Codex so agents can continue tasks for hours or days in customers' clouds.