What Is a Search Center
A search center is the same underlying assistant, served in search mode: a public, full-screen page whose main element is a search box over one dataset. Every answer is grounded in your captured content and shows the documents it came from. You build it from a memory dataset exactly like an assistant — what differs is the experience it presents and where it lives.
Search Center vs Assistant
Both answer from a dataset with citations. Choose based on the experience you want:
| Search center | Assistant | |
|---|---|---|
| Surface | A full-page, standalone search page | A floating bubble or inline widget on your site |
| First impression | A search box and suggested queries | A chat greeting and a conversation |
| Best for | A public "search our content" destination — docs, catalog, knowledge base | In-context help alongside your existing pages |
| Access | Public by default, with its own shareable link | Public or private, embedded in your pages |
| History | Each visit starts fresh (by default) | Remembers the conversation |
You can run both from the same dataset — a search center as the public destination, and an assistant embedded across your site.
When to Use a Search Center
- You want a dedicated search page people can bookmark and link to.
- Your visitors think in queries, not conversations — documentation, catalogs, help centers.
- You want a public, no-login search over your content that you can also embed.
Related Docs
- Core concepts — how datasets, assistants and answers fit together.
- Build a Perfect Assistant — the conversational counterpart.