Claude Artifacts vs Projects: what is the difference?

Short answer

An Artifact is an output: a document, deck, dashboard or small tool that Claude makes and opens beside the chat, saved in the Artifacts tab and shareable. A Project is an input: a workspace whose knowledge, instructions and memory every chat inside it starts with. You often use both: a Project supplies the context, and the Artifacts are what that context produces.

Output versus input

The two features sit at opposite ends of a conversation. A Project is what Claude starts with: the knowledge files, the instructions, the memory of that workspace. An Artifact is what Claude hands back: in Anthropic’s words, content Claude makes that you would put in front of someone, such as a design, deck, document, dashboard or small interactive tool.

Project Artifact
Role Context going in Result coming out
Contains Knowledge, instructions, chats, memory One document, deck, dashboard, design or tool
Where you find it Projects in the sidebar Artifacts tab in the sidebar
Sharing Team and Enterprise, Can view or Can edit Starts private; can be shared
Plans All; Free up to five Projects All plans in chat; some templates Pro and above
Needs Nothing extra “Create and edit files” (code execution) switched on

When Claude makes an Artifact

Claude typically creates an Artifact when the content is longer than about 15 lines, self-contained and likely to be reused. It opens beside the conversation, where you edit it by asking Claude or, in templated Artifacts, directly. You can also start from a template: Docs, Slides, Design, Data, Sheets, Whiteboards and Motion, with different plan availability. Template Artifacts export to formats like Word, PDF, Markdown, PowerPoint, Google Slides, standalone HTML or MP4.

Artifacts can also do things a Project cannot. On Pro and above they can connect to apps you linked to Claude and store data between sessions, either personal (each user sees their own) or shared (everyone sees the same). Artifacts that use Claude let other people use the tool with their own Claude account, so the usage counts against their plan, not the creator’s.

When you need a Project instead

If the question is “how do I get Claude to remember our product details in every chat”, an Artifact is the wrong tool. Artifacts do not feed context into new chats. That is the job of a Project: its instructions apply to every chat in it, and its knowledge is read or searched in each one.

A common pattern for a team:

  1. Make a Project per client or product, with instructions and the reference files.
  2. Work in chats inside that Project.
  3. Let Claude produce the deliverables as Artifacts: the proposal doc, the slide deck, the dashboard.
  4. When a deliverable becomes a reference for later work (the approved proposal, the final price table), add its content to the project knowledge so future chats build on it.

The misunderstanding: an Artifact is saved into the Project

People assume that an Artifact made inside a Project becomes part of that Project’s knowledge. Anthropic’s artifact documentation does not describe such a link: Artifacts live in the Artifacts tab, and project knowledge is what you add with the “+” on the Project page. If the approved version should guide the next chat, put it in the knowledge yourself. Otherwise the next chat works from the older files.

What this means for a team

Sharing an Artifact shares a result: a colleague can open the deck or use the tool. Sharing a Project shares the context: colleagues can chat with the same knowledge and instructions. Neither shares what each person taught Claude along the way; that stays in each person’s memory. And neither travels to ChatGPT or Copilot. For more on the context side, see how to use Claude Projects effectively and Claude Skills vs Projects.

Where Bombos fits

Artifacts are where the work ends up; Projects are a hand-filled folder of what Claude should know. Bombos is a plugin that gives every AI tool your team uses, Claude and Claude Code as well as ChatGPT, Copilot, Cursor and your own agents, one shared memory that reads what already lives in mail, documents and repos and keeps the rules your team learned, so the approved version is known to the next tool without someone copying it over. One person can try it first for their own agents.

Related questions

Where do I find my Artifacts?

All Artifacts are saved in the Artifacts tab in the sidebar. On iOS and Android you can view them there too, but sharing changes requires the web or Desktop app.

Do Artifacts need a setting turned on?

Yes. Artifacts need "Cloud code execution and file creation", called "Create and edit files" in the updated layout, under Settings > Capabilities or, on Team and Enterprise, Organization settings > Capabilities.

Can an Artifact use my data?

On Pro and above, Artifacts can read from and write to apps you connected to Claude, and can store data, either personal per user or shared by all users.

Can I put an Artifact into a Project?

Anthropic's artifact article does not describe that. You can add the content as a file to project knowledge yourself, so later chats in the Project can use it.

Sources

  1. What are artifacts and how do I use them? (Claude Help Center)
  2. What are projects? (Claude Help Center)
  3. How can I create and manage projects? (Claude Help Center)
The first half hour · free

What does your AI forget?

  1. 01You tell us which AI you use and where it keeps starting over.
  2. 02We show what a shared memory changes there, on your own examples.
  3. 03You keep the plan, with or without Bombos.
Piet Baudoin
Piet BaudoinYou talk to Piet, from the Bombos teamDutch team from Utrecht and Nijmegen · data in Europe
We reply the same working day.