Claude Artifacts: Create, Share and Publish (2026 Guide)

Quick answer: Claude artifacts are standalone pieces of content Claude creates beside your chat: documents, code, single-page websites, diagrams, dashboards, slide decks or small interactive apps. You can edit them, find them in the Artifacts tab, and publish or share them by link. They’re available on every Claude plan, including Free, and in Claude Code.

Artifacts started as a side panel for code and documents. In 2026 they’re closer to a lightweight app platform: templates for docs, slides and designs, AI-powered apps that call Claude, storage, and live data from connectors. This guide covers what you can do today and where the limits are, checked against the Claude Help Center in October 2026.

What counts as an artifact

Anthropic’s definition: an artifact is anything Claude makes for you that you’d put in front of someone. In practice:

• Documents (Markdown or rich text), reports, specs
• Code snippets and small programs
• Single-page websites (HTML)
• SVG images, diagrams and flowcharts
• Dashboards and data visualizations
• Slide decks and visual designs
• Small interactive tools and games

Claude creates one on its own when content is significant and self-contained, typically over 15 lines, and something you’ll likely edit, iterate on or reuse. For a quick answer or a short explanation, it stays in the chat.

How to turn artifacts on

Artifacts rely on Claude’s code execution environment. If Claude never opens one, check this setting first:

Plan	Where to enable
Free, Pro, Max	Settings > Capabilities > Cloud code execution and file creation (shown as Create and edit files in the updated layout)
Team, Enterprise	Organization settings > Capabilities (an Owner controls it)

How to create and edit an artifact

1. Ask for it. “Build an interactive budget calculator”, “turn these notes into a one-page brief”, “make a flowchart of our deploy process”. Saying “make this an artifact” works when Claude doesn’t create one by itself.
2. Or start from a template. Select Output in the message box and choose a template.
3. Iterate. Ask Claude for changes in the chat. On template-based artifacts you can also edit directly, or highlight a passage and use Edit with Claude for a targeted change.
4. Branch. Editing an earlier message creates a separate version, so you can explore a different direction without losing the first one.
5. Find it later. Everything you create is saved in the Artifacts tab in the sidebar, across conversations.

Artifacts also work on the iOS and Android apps: ask for a design, deck or doc in any chat and open it from the Artifacts tab.

Clear prompts produce better artifacts. Our Claude prompting best practices apply here too: say who the artifact is for, what it must include, and what “done” looks like.

Templates

Templates give artifacts a dedicated editor and export options:

Template	What it’s for	Exports
Docs	Collaborative living documents	Word, PDF
Slides	Presentations	PowerPoint, PDF
Design	Mockups, landing pages, screen flows	PNG, PDF per artboard
Dashboards (beta)	Data visualizations, including from connected apps	—
Motion (beta)	Animated explainers	MP4

Availability and defaults vary by plan and change often. As of October 2026 the Help Center lists Dashboards in beta on Pro, Max, Team and Enterprise, and Motion in beta on Team and Enterprise, with Enterprise Owners turning some templates on themselves. Check the Help Center for your plan.

Publishing, sharing and remixing

Artifacts start private to you. How you share depends on your plan:

	Free, Pro, Max	Team, Enterprise
Button	Publish	Share
Default audience	Public link	Your organization
Options	Public link, or back to only you	Only people invited, anyone at your organization, or anyone with the link
Roles	—	Viewer, commenter or editor (not every artifact type supports all three)
Admin control	None	Admins can disable external sharing, which stops public links working

A few rules worth knowing before you send a link:

• Depending on the artifact type, viewers may need a Claude account to open it. Artifacts that use Claude or connect to apps can’t use “anyone with the link” on Team and Enterprise.
• You can invite up to 50 external people per artifact.
• Unpublishing is one-way. Anthropic notes you can’t publish an artifact again after you unpublish it, and its stored data is deleted.
• Remixing: signed-in users can copy a published artifact and build on it. The copy is separate, so nothing they do affects your original.

AI-powered artifacts

An artifact can call Claude itself, which turns it into a small AI app: a tutor, a writing coach, a quiz, a game master, a support-ticket classifier. People using your shared artifact sign in with their own Claude account and get their own instance.

The economics are unusual. No API key is required, there’s no cost to you, and usage counts against each viewer’s own plan limits, whether ten or ten thousand people use it. If you outgrow that model, for example you need your own backend, billing or custom auth, move to the Claude API and a real deployment.

Storage and connectors

On paid plans, artifacts can go beyond a static page:

• Storage. An artifact can keep data, either personal (private to each user) or shared (visible to everyone who uses it). The limit is 20 MB per artifact and text only: no images, files or binary data.
• Connectors. Artifacts can pull from connected apps such as Asana, Google Calendar and Slack, plus custom connectors. The first time an artifact needs a connected app, Claude shows which apps and tools it will use. Connectors are built on the Model Context Protocol.

On Team and Enterprise, Owners can switch artifact connectors off for the whole organization.

Artifacts in Claude Code

Claude Code can publish artifacts too. It turns session output into a live web page at a private claude.ai URL that updates in place as the session continues. Good uses:

• A pull-request walkthrough with annotated diffs
• A dashboard built from data the session pulled
• Several design or implementation options side by side
• A progress checklist that fills in during a long migration

Ask in plain language, for example “make an artifact that walks through this PR with the diff annotated inline”. Run /artifacts to list artifacts you own or that were shared with you.

Requirements, per the Claude Code docs: a Pro, Max, Team or Enterprise plan, a session signed in with /login (API key and cloud-provider sessions can’t publish), and the Anthropic API as model provider. Each Claude Code artifact is one self-contained page with no backend, and the rendered page must be 16 MiB or smaller.

Artifacts vs Projects vs Claude Code

These three are easy to confuse:

Feature	What it holds	Best for
Artifacts	Outputs Claude creates	Documents, apps, decks, dashboards you’ll edit or share
Projects	Inputs: files and instructions reused across chats	A knowledge base or recurring workflow
Claude Code	Your real codebase	Building and changing software in a repository

A common pattern: keep reference docs in a Project, ask for an artifact inside that Project, and graduate the artifact to a real repository with Claude Code once it needs a backend.

Limits to keep in mind

• No backend. Artifacts are pages, not servers. They can’t run your own server code or authenticate users themselves.
• Text-only storage, capped at 20 MB per artifact.
• Plan-gated extras. Connectors and storage need a paid plan; some templates are beta and plan-specific.
• Usage counts. Creating and iterating on artifacts uses your plan’s limits like any other message. Our Claude pricing guide explains the plans, and Claude Pro vs Max helps if you hit limits often.

For a wider view of the apps and features, start with What is Claude?, or compare the experience with ChatGPT in Claude vs ChatGPT.