Claude Artifacts: What They Are, What You Can Build, and 7 Artifacts for Operators

An artifact is the moment Claude stops talking and hands you something you can put in front of a client.

Most operators use AI as a conversation. They ask, it answers, they copy the answer somewhere else and format it. The output is text, and text is the least leveraged form a deliverable can take.

Artifacts change the unit of output. Instead of a paragraph describing a pricing calculator, you get the calculator. Instead of advice about an intake form, you get the form. The work arrives in the shape you would have built by hand, ready to edit, share or publish.

This article explains what artifacts are according to Anthropic's own Help Center and announcements as of October 2026, how templates, sharing, storage and connectors work, how artifacts differ from Projects and Skills, and then gives seven artifacts a coach or owner-operator can build in a single session, with the prompt to ask for each one. Product details change; the Help Center article on artifacts is the page to check.

What an artifact is, in Anthropic's words

The Help Center definition is the clearest one available: "An artifact is anything Claude makes for you that you'd put in front of someone: a design, a deck, a document, a dashboard, or a small interactive tool. It opens beside your conversation, and you can edit it, come back to it, and share it with others."

Claude decides when to produce one rather than a plain reply. The same page lists the criteria:

• "It's significant and self-contained, typically over 15 lines."
• "It's something you're likely to want to edit, iterate on, or reuse outside the conversation."
• "It stands on its own without needing extra context from the conversation."
• "It's content you're likely to refer back to or use later."

The range is wide. "Claude can make an artifact out of almost anything, including documents, code snippets, single-page websites, images, diagrams and flowcharts, dashboards, and small interactive tools."

Availability, as of October 2026: "Artifacts are available on Free, Pro, Max, Team, and Enterprise plans, and in Claude Code on every plan that includes Claude Code." There is one switch to know about. The page states that artifacts "require Cloud code execution and file creation to be turned on in Settings > Capabilities" on individual plans, or in organization settings on Team and Enterprise. If artifacts never appear for you, that setting is the first place to look.

One date matters for anyone who built with artifacts earlier. The Help Center notes: "Legacy artifacts are artifacts made in a chat before September 16, 2026. They keep working, and you can still publish and share them, but you can't make new ones." The newer system described below replaced them.

Templates: Docs, Slides and Design

Artifacts can start from a template rather than a blank conversation. The Help Center lists three: Docs, Slides and Design, and notes they are "in beta on paid plans only." They are on by default on Pro, Max and Team, and off on Enterprise until an owner enables each one.

The descriptions are specific enough to plan around:

• Docs: "Living documents you write with Claude and your team in real time."
• Slides: "Presentations built from your notes, reports, or the work already in your chat. Edit any slide directly, present without leaving Claude, and export to PowerPoint or PDF."
• Design: "Visuals, mockups, prototypes, one-pagers, and landing pages, built with your design system."

Export formats are stated plainly: "Docs export to Word, PDF, Markdown, and Google Docs. Decks export to PowerPoint and PDF." Designs export "as a .zip, PDF, PowerPoint, or standalone HTML". For an operator this means the proposal, the deck and the landing page mockup can all leave Claude in the format the client expects.

Editing works two ways. You can ask Claude to change the artifact, or edit directly inside it. In a document you can "highlight the text you want changed, click 'Edit with Claude,' and type your request", which is faster than describing the paragraph you mean.

Sharing and publishing

The default is private. The Help Center article on sharing opens with: "Artifacts start private to you. Nothing is shared until you share it."

On Pro and Max the choice is "Only you" or "Anyone with the link", plus an email invitation feature described as "Available in beta on Pro, Max, Team, and Enterprise plans." Invitations carry an access level: "Can view", "Commenter" or "Can edit". "Pending invitations expire after 30 days. Accepted invitations don't expire."

Teams run under tighter rules. "On Team and Enterprise plans, artifacts stay inside your organization by default. An owner decides whether users can share outside it." The admin guide describes two separate switches, External sharing and Email invitations outside your organization, and warns that turning External sharing off "also stops existing public links from working until you turn it back on".

Two limits on invited outsiders are worth knowing before you promise a client a working tool. "Parts of the artifact that use Claude or your connectors don't work for them." And people invited from outside your organization "can't invite others or change sharing settings."

Artifacts that use Claude, connect to apps and store data

This is where artifacts stop being documents and become small applications.

Artifacts that call Claude

The Help Center describes it directly: "You can build artifacts that call Claude directly, turning them into small apps." The cost model is the part operators care about: "No API keys are required, and there's no cost to you. Whether your artifact helps 10 people or 10,000, sharing is free, and usage counts against each person's own plan limits rather than yours."

Anthropic's July 2025 announcement of this capability listed the same economics as bullet points: "They authenticate with their existing Claude account", "Their API usage counts against their subscription, not yours", "You pay nothing for their usage", and "No one needs to manage API keys". The earlier June 2025 post gave the scale at that point: "millions of users have created over half a billion artifacts".

For a coach, this means a client-facing tool that asks questions and adapts its answers, shared by link, with no server, no API bill and no deployment.

Connect your apps

"Available on Pro, Max, Team, and Enterprise plans, on Claude on web and desktop." Artifacts "can read from and write to tools like Asana, Google Calendar, and Slack", and "They can also connect to any custom connectors you've set up." Approval is explicit: the first time an artifact needs a connected app, Claude shows which tools it will use and asks. "Everyone connects their own apps, even when using a shared or published artifact."

Store data

Artifacts "can store data between sessions, so you can build things like journals, trackers, and collaborative tools." Storage is personal or shared: "In a journal artifact, your entries stay visible only to you." "In a game leaderboard, everyone sees the same scores." The cap is specific: "Storage has a 20 MB limit per artifact and accepts text only, no images, files, or binary data."

The Help Center's warning deserves repeating to any client you hand a tool to: "Before entering anything sensitive, check whether the artifact uses shared storage."

Artifacts versus Projects versus Skills

These three get confused because they all make Claude more useful. They do different jobs.

• A Project is a workspace. It holds instructions and files so Claude knows your business before you type. The Claude Projects guide covers the setups a coaching business needs.
• A Skill is a procedure. It is a written process Claude follows when the task matches, so you stop re-briefing. The Claude Skills guide explains the format.
• An artifact is an output. It is the thing produced: the document, deck, dashboard or tool, and it can be shared or published on its own.

The compound version uses all three. A Project holds your offer, pricing and client language. A Skill holds your onboarding procedure. The artifact is the onboarding checklist that comes out, formatted and shareable, in one request.

Seven artifacts an operator can build in one session

Each item below is a tool a coach, consultant or owner-operator uses weekly. Ask for it inside a Project that already holds your offer and pricing, and the first draft will be close. Then iterate in plain language, as the Help Center suggests: "make the buttons bigger", "change the color scheme", and so on. The prompts are written in a form you can paste.

1. Client intake form mockup. A single-page form with the fields your discovery call needs. Prompt: Build an interactive intake form for a performance coaching practice: name, email, main goal, training history, schedule constraints, biggest obstacle, and a 1 to 10 readiness score. Show a clean summary card after submit.
2. Pricing calculator. A tool that turns your inputs into a quoted price. Prompt: Build a pricing calculator for coaching packages. Inputs: number of sessions per month, programme length in months, group or one-to-one. Output: monthly price, total price, and the per-session rate, using the pricing in this Project.
3. Weekly review dashboard. A page that captures the Stoic weekly review and shows the trend. Prompt: Build a weekly review tracker with personal storage. Each week: what went well, what went badly, what changes next week, plus 1 to 5 scores for training, sleep and focus. Show the scores as a 12-week chart.
4. Habit tracker. Daily checkboxes for the keystone habits, with streaks. Prompt: Build a daily habit tracker with personal storage for five habits: training, morning review, no caffeine after noon, 10,000 steps, lights out by 10:30. Show current streak and a 30-day grid.
5. Offer one-pager. A Design-template one-pager for the programme you sell. Prompt: Using the Design template, create a one-page overview of the 90-day programme in this Project: headline, who it is for, the three pillars, what is included, price, and a call to action. Keep it to one printable page.
6. Onboarding checklist. A shared checklist the client ticks off in week one. Prompt: Build an onboarding checklist with shared storage for a new coaching client: contract signed, intake form, baseline measurements, app installed, first session booked, first training week logged. Each item shows who completed it and when.
7. Landing page wireframe. A grey-box layout to hand a developer or build from. Prompt: Using the Design template, wireframe a landing page for the free 5-day challenge: hero with headline and email field, three benefit blocks, a short how-it-works section, one testimonial slot, and a final call to action. No colour, structure only.

Two of these, the tracker and the onboarding checklist, depend on storage, which the Help Center lists as available on Pro, Max, Team and Enterprise. The one-pager and wireframe depend on the Design template, which is in beta on paid plans. On Free, build the other three first.

If you want the client-facing versions to think, ask Claude to add a question-and-answer step and the artifact will call Claude for whoever opens it, at their cost, not yours. That is the same mechanism behind the lead-magnet approach in how to build a lead magnet with Claude.

Frequently asked questions

What are Claude artifacts?

Per Anthropic's Help Center, an artifact is anything Claude makes for you that you would put in front of someone: a design, a deck, a document, a dashboard or a small interactive tool. It opens beside the conversation and can be edited, revisited and shared. Claude creates one when the content is significant, self-contained, typically over 15 lines, and likely to be reused.

Are Claude artifacts free?

As of October 2026 the Help Center lists artifacts as available on Free, Pro, Max, Team and Enterprise plans. Templates for Docs, Slides and Design are in beta on paid plans only, and storage and app connections are listed for Pro, Max, Team and Enterprise. Artifacts that call Claude cost the builder nothing; usage counts against each user's own plan.

How do I share a Claude artifact?

Open the artifact and click Share. On Pro and Max you choose between only you and anyone with the link, or invite specific people by email with view, comment or edit access. On Team and Enterprise plans artifacts stay inside the organization by default, and an owner decides whether external sharing is allowed. Sharing settings cannot be changed from the mobile apps.

Can Claude artifacts store data?

Yes, on Pro, Max, Team and Enterprise plans on web and desktop. Storage is either personal, where each person sees only their own entries, or shared, where everyone sees the same data. The limit is 20 MB of text per artifact, with no images or files. Check whether a tool uses shared storage before entering anything sensitive.

What is the difference between artifacts and Projects?

A Project is a workspace holding instructions and files so Claude starts with your context. An artifact is an output, the document, deck or tool Claude produces, which can be shared or published on its own. Skills are a third thing: written procedures Claude follows when a task matches. Used together, the Project supplies context, the Skill supplies the method, and the artifact is the result.

Ask for the thing, not the description of the thing

The shift artifacts make is small to describe and large in practice. You stop asking Claude to tell you about a checklist and start asking for the checklist. The deliverable arrives built, you edit it in place, and you hand it over by link.

That is leverage in the operator's sense: the same hour producing a finished tool instead of notes toward one. It only compounds if the rest of your week is built the same way, with training anchored, reviews written and the day structured before it starts. The free 5-Day Stoic Operator Challenge installs that structure in five days. Build your first artifact on day one; by day five you will know which of the seven above your business actually needs.