Claude Artifacts is a feature in Claude, Anthropic’s AI assistant, that opens a separate window next to the chat and puts substantial content there — a piece of code, a webpage, a document, a diagram — instead of burying it in the conversation. You can keep talking to Claude to revise it, and the changes update live in that window. In short: it turns a conversation into something you can look at, edit, and hand to someone else.
What it is
When you ask Claude to build something concrete — a small web app, a chart, a slide-style document, a diagram — Claude often creates an artifact rather than pasting a giant code block into the chat. Anthropic’s own guidance is that this happens for content that is “significant and self-contained” and something you’re likely to edit, reuse, or refer back to, typically anything longer than about 15 lines. The artifact appears in its own panel, rendered and interactive where possible, while the conversation continues on the left.
Artifacts can hold several kinds of content: formatted documents (Markdown or plain text), code in most common languages, single-page HTML sites, SVG images, diagrams and flowcharts, and interactive React components. That range is what makes the feature useful beyond coding — it works just as well for a one-page pitch document or a data visualization as it does for a script.
How it works
You don’t do anything special to create one — you just describe what you want in normal conversation, and Claude decides whether the result belongs in an artifact. If you want to change it, you ask Claude for the edit in the chat, and the artifact panel updates in place rather than Claude re-pasting the whole thing. Because the content lives outside the scrolling conversation, it’s easier to keep iterating on a single working version instead of hunting through chat history for the latest copy.
Artifacts don’t save to your account automatically. To keep one and make it reusable, you open it and click “Publish,” which adds it to your artifacts collection and, if you choose, makes it shareable — either privately with people who have the link or, on the Team and Enterprise plans, with your whole organization, with named collaborators able to edit it together. Some artifacts can also go further and embed Claude’s own intelligence, so that anyone who opens a published, AI-powered artifact can interact with it — ask it questions, get a personalized response — without needing their own API key. Their use counts against their own Claude usage, not the creator’s account. Artifacts can also connect to outside tools like calendars or messaging apps through MCP, Anthropic’s protocol for linking AI systems to external data and services, which lets an artifact read from or write to a service the user has connected.
Why it matters
Most chatbot conversations produce a wall of text that’s hard to reuse. Artifacts turn that text into an object — a document, a mini-app, a visual — that persists, can be edited without starting over, and can be handed to a colleague as a working thing rather than a copy-pasted answer. That distinction matters for anyone using Anthropic’s AI for real work: a marketer building a one-off landing page mockup, a teacher generating a worksheet, an analyst producing a chart from a dataset, or a developer prototyping a small tool, all get a result they can keep iterating on and pass along, instead of a message that scrolls out of view.
How to get started
Artifacts are available on all Claude plans, including the free tier, but the underlying capability has to be switched on: in Settings → Capabilities, turn on “Code execution and file creation” (organization admins do the equivalent in workspace settings for Team and Enterprise accounts). After that, just ask Claude to build or write something substantial in a normal conversation — no special command is needed — and Anthropic’s own walkthrough on artifacts covers the details, including publishing and sharing.
In the news
Anthropic recently expanded what teams can do with the feature, adding shared, team-editable artifacts and simpler public links, including the ability to create one directly from a Slack channel using Claude Tag.
FAQ
Does an artifact cost anything to share? No — sharing is free, and if the artifact itself uses Claude’s intelligence, the people using it draw on their own usage limits, not the creator’s.
Can I edit an artifact after it’s created? Yes. Ask Claude for the change in the same conversation, and the artifact panel updates instead of Claude repeating the whole thing in chat.
Is this the same as Claude Code? No. Claude Code is a separate, developer-focused coding tool; Artifacts is a chat feature for turning any substantial output — not just code — into a standalone, shareable piece of content.
Do I need to know how to code to use it? No. You describe what you want in plain language; Claude writes the underlying code or content.
Sources: Claude Help Center — What are artifacts and how do I use them?