Claude Code Artifacts Enterprise-Only? Use PandaNpc to Let Anyone Publicly Share AI Coding Results

Anthropic launched Artifacts for Claude Code: it turns the output of a coding session into a live, interactive webpage you can send to colleagues as a link. That's a great direction—the long string of terminal output was never a good format for sharing with others. But when you actually try to use it, you'll find it shuts out most developers: it's only available on Team / Enterprise plans, links can only be opened by members of the same organization, and the page is read-only.

If you're an individual subscriber, run Claude Code on a self-hosted backend, or simply want to share your work in a group chat with a friend outside your organization—the official Artifacts won't help. This article first explains the features and limitations of Claude Code Artifacts, then shares an open, accessible-to-everyone alternative: PandaNpc result pages.

1. What Claude Code Artifacts is and what it can do

Simply put, an artifact is a self-contained HTML page that Claude Code publishes from your session, hosted on a private URL at claude.ai. As the session continues, the same link refreshes in place. It's not an export of the chat log, but a results page “designed for reading and interaction.” According to the official docs, it can contain:

• Annotated change walkthrough: renders a PR's diff, adds comments next to relevant code lines, and color-codes by severity, so reviewers can follow your thinking instead of piecing it together themselves.
• Option comparison: place several implementation / layout / copy variants side by side on one page, with a one-line tradeoff under each.
• Interactive tuning: sliders and toggles bound to a value (such as an animation easing curve) with live preview.
• Dashboard / investigation timeline: generate charts, metric cards, or a timeline that fills in as you go from session data.
• Bring results back to the session: a “Copy as prompt” button on the page exports the interactive result as text so you can paste it back into the terminal and let Claude continue.

Combined with Claude's built-in design skills (automatic color palettes, layout, and priority use of your project's design system), these capabilities really make “showcasing results” much more presentable.

2. Three limits: why most people can't use it

After going through the features, now look at the hard requirements explicitly stated in the official docs—and here's the problem:

1. Team / Enterprise plans only: you also must log in to claude.ai via /login. If you run Claude Code with an API key, gateway token, Bedrock / Vertex, or any other method, you cannot publish artifacts.
2. Only members of the same organization can view it: sharing stops “at the edge of your organization.” People outside the organization—clients, external collaborators, friends in a WeChat group—would need to log in to the same organization to open the link, otherwise they can't see it. The only “external” method the official docs offer is manually sending the HTML file itself to the other person.
3. Read-only, no collaboration: other people can only view the version you published; they can't respond or continue on the page.

In other words, Claude Code Artifacts is an enterprise internal collaboration feature. For individual subscribers—the vast majority of developers—small teams, and people using custom backends, the door is basically closed.

3. PandaNpc result pages: public, available to everyone, and can be brought back into the session

PandaNpc's approach is exactly what opens that door the official feature closed: it turns “sharing session results” into public links + available on any plan + viewable without login. When you share a Claude Code session in PandaNpc, check “Generate result page,” and the backend automatically reads the session and generates a structured result page, hosted on your own public link (pandanpc.com/share/cc/...). Anyone can open and view it without needing a Claude account or belonging to any organization.

What's on the result page (each block appears automatically depending on whether the session has matching content):

• One-sentence summary (TL;DR): what this session actually accomplished.
• Session scale bar: time spent, number of conversation turns, tool-call distribution (Bash×N, Edit×N…), and how many files were created/modified—see the scope of the work at a glance.
• What was done: a bulleted list.
• Changed files + real diff cards: lists which files changed; click to reveal a syntax-highlighted red/green diff. This is the open version of the official “annotated change walkthrough,” bringing the result page up to “good enough to send to a colleague as a PR walkthrough.”
• Commands run: lists the commands this session ran (e.g., npm test) in terminal style.
• Problem → Solution: the problem this session faced and how it was solved.
• Remaining items: what to-dos are still unfinished.
• Tags: bugfix / feature / refactor + the tech stack involved, for easy categorization and retrieval.
• Copy as prompt: one click assembles the result summary into a prompt and copies it—this is a pure front-end version of the official “Copy as prompt (bring results back to the session)” and works for anyone who opens the page.

At the bottom you can also expand the full session replay, so anyone who wants to see the whole process can browse through it anytime.

4. How to use it: three steps to turn a session into a public result page

1. Open a Claude Code session in PandaNpc (on desktop or the web /agent).
2. Click Share → check “Generate result page” (on by default) → generate the link. The backend uses a model to read the session, generates the result summary in a few seconds, and caches it.
3. Send out the public link. The recipient sees the result page directly without logging in or belonging to any organization; to see details, they can expand the full replay.

If the session later sees more progress, go back to the share dialog and click “Update snapshot,” and the result page will be regenerated based on the latest content.

5. PandaNpc result pages vs official Artifacts

6. Which scenarios it fits

• Deliverables for clients / external collaborators: people outside your organization can open it; no need to bring them into your Claude org.
• Team review: diff cards + problem/solution are much clearer than pasting a chunk of terminal output.
• Knowledge capture and retrospectives: tags + changed files + commands make it easy to later search for “how did we fix that one.”
• Personal sharing: you don't need a Team / Enterprise plan to send out a polished result page.

Conclusion

Claude Code Artifacts points in the right direction—turning AI coding results into shareable webpages—but the enterprise plan requirement and organization-only visibility confine it to an internal tool. PandaNpc result pages preserve the same value—turning a session into a clean results showcase—while making it public, available to everyone, and one-click returnable to the session. No matter which plan or backend you use to run Claude Code, you can share your work openly.

Want to share more than just result pages? You can also securely share your Claude subscription with friends or teammates to use together—they remotely use the Claude Code on your machine, without your account password, and you can revoke access anytime with one click.