Skip to content

Claude Code Artifacts Go Live and Shareable: Agent Output Becomes a Dashboard Your Team Can Open

July 15, 2026. Anthropic spent the last two weeks shipping ways to delegate work to Claude. On July 14 it shipped the other half: a way to hand the results to people. Artifacts, the interactive pages Claude builds, now work inside Claude Code, publish as live pages a whole team can open, and can be shared publicly or embedded outside your workspace.

What Anthropic shipped

Per Anthropic's announcement and the Claude Code documentation, with independent coverage from VentureBeat and The Decoder, the July 14 update includes four pieces.

  1. Artifacts in Claude Code (beta). A coding or analysis session can now publish its output as a live web page. When Claude produces a new iteration, the open page refreshes in place and holds the viewer's scroll position, so teammates watch updates land in real time. In beta for Team and Enterprise plans.
  2. Public sharing and embeds. Published artifacts can be shared to public links viewable without signing in, and users on Free, Pro and Max plans can generate an embed code for any published artifact.
  3. Governance defaults that matter. On Team and Enterprise plans, external sharing is off by default. Members share within the organization only, until an Owner enables External sharing under Settings, Claude Code, Capabilities.
  4. Builds from Slack via Claude Tag. Artifacts can be created and edited through Claude Tag, the Slack-native agent we covered at launch, so a dashboard request can start as a channel message and end as a shared page.

What it means for operators

The unglamorous truth about agent work is that the output usually dies in a chat transcript or a screenshot. This update gives the output a URL. For agencies and SMB teams, three patterns are immediately practical. Client reporting: a live page the client can open any time beats a Friday PDF, and an agent can keep it current as part of a scheduled run, the delegation pattern from Cowork's cross-device rollout. Internal ops: pipeline trackers, QA boards and campaign monitors that a Claude Tag request can spin up without a developer sprint. Deliverable prototypes: an interactive mockup shared as a public link closes discovery calls faster than a slide deck.

Two cautions before you wire this into client work. The external sharing default is off for a reason: check what a page exposes (data, client names, internal metrics) before an Owner flips the toggle, and treat public artifact links like any other unauthenticated URL. And live artifacts are a presentation layer, not a data platform: for dashboards that need real pipelines, warehouse connections and access control, this complements rather than replaces proper AI automation infrastructure. If you want agent-built reporting wired into your stack properly, an experienced AI engineer can set up the pattern in a week: agents do the work, artifacts show the work, humans approve the work.

Want agent-built dashboards running your ops?

We design, build, and run it for you, integrated with the tools you already use. Free audit in 24 hours.

Get Your Free Audit

Frequently Asked Questions

Artifacts are interactive pages Claude generates, and as of Anthropic's July 14, 2026 update they work inside Claude Code: a session can publish its output as a live web page that refreshes in place as Claude iterates, holding the viewer's scroll position. Teammates open the page and watch updates land in real time. The capability is in beta for Team and Enterprise plans.

Yes, published artifacts can be shared to public links that anyone can view without signing in, and Free, Pro and Max users can generate embed codes for published artifacts. On Team and Enterprise plans, external sharing is off by default: members can share only within the organization until an Owner turns on External sharing under Settings, Claude Code, Capabilities.

Three quick wins: live client reporting pages an agent keeps current on a schedule instead of static weekly decks, internal ops dashboards like pipeline trackers spun up from a Slack message via Claude Tag, and interactive prototypes shared as public links during sales conversations. The pattern is agents do the work, artifacts make the work visible, humans approve it.

No. Artifacts are a fast presentation layer for agent output, ideal for reports, trackers and prototypes. Dashboards that need governed data pipelines, warehouse connections, row-level access control and uptime guarantees still belong in proper BI or custom infrastructure. Use artifacts to ship visibility in hours, and graduate the ones that prove valuable.

Free Strategy Audit

Ready to put this to work?

Join 200+ businesses already scaling with AI and automation. Get your free audit and a custom roadmap within 48 hours.

Website & marketing performance analysis
AI & automation opportunity mapping
Custom growth roadmap with ROI estimates
Delivered within 48 hours, 100% free
200+
Clients served
48hr
Turnaround
100%
Free, no strings

Get Your Free Audit

Takes 30 seconds. No credit card required.

Prefer to chat?

WhatsApp us