Rule of thumb: if you'd want it in the next matter too, it belongs in a database. If it only makes sense inside this one, leave it in Project files.
## Before you start
Databases are part of the **Team** plan, and your firm administrator has to switch them on. If your plan includes them but they haven't been enabled yet, ask your administrator. If your team is below the Team plan, **Databases** appears with an upgrade prompt in place of the list. See [Subscription](/en/resources/subscription).
When databases are available to you, **Databases** appears in the global part of the sidebar, next to **Home** and **Templates** — above the project switcher, because it doesn't belong to any one project.
## The two kinds
| Kind | Where the files live | Use it when |
| ----------------------- | ------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- |
| **Libra database** | Uploaded into Libra. | You're building a set from files on your computer, or from documents already in a project. |
| **SharePoint database** | Left in SharePoint; Libra links to the folder and keeps it in sync. | The authoritative copy should stay in SharePoint and stay current. Requires the [SharePoint integration](/en/integrations/sharepoint). |
Both behave the same once built. The difference is only where the files are stored and who keeps them up to date.
## What's inside one
Opening a database shows its documents in a browser you can sort by name, author, or date added. Each row carries the document's **author**, when it was **added**, and your **access** to it. Folders behave as first-class items: you can select, move, and delete a whole folder and everything under it.
You can fill a database by uploading files or entire folders, or by pulling in documents that are already in a project's [Project files](/en/features/documents/library) — those are linked, not re-uploaded, so there's only ever one copy.
## Using a database
A database is a source you can point Libra at from anywhere:
| Where | How it works |
| ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Chat** | Attach the database from **Tools → Search files**, or `@`-mention it. Libra searches it on demand rather than loading every file, so a large database stays fast. |
| **Review** | Run a Playbook against documents from the database. See [Reviews](/en/features/review/basics). |
| **Discovery** | Extract the same columns across every document in the database. See [Discovery](/en/features/discovery/basics). |
Because the database is shared, the answer a colleague gets in another matter is drawn from exactly the same sources as yours.
## Sharing and access
The person who creates a database owns it and controls sharing. When a database is shared with you, your role on it — **Viewer**, **Editor**, or **Admin** — is shown in the browser and decides whether you can add or remove documents.
## Your page allowance
Uploading into a **Libra database** draws on a page allowance that scales with your subscription: each billed seat grants **100,000 page-equivalents**, pooled across the team.
Two things to know about it:
* It's a **total, not a monthly reset.** The allowance covers everything your team has ever ingested into Libra databases.
* **Deleting documents doesn't give pages back.** If you're close to the limit, add seats or buy more pages rather than clearing space.
Documents in a **SharePoint database** don't count toward it, because Libra doesn't ingest them. If an upload would take you past the allowance, Libra refuses it and tells you how many pages are left.