Data sources

Requirements

Data sources are available starting with the Business license. With the Starter or Pro license the area is not accessible.

Storage

Everything 9brains makes searchable for you shares one pool of storage: the connected data sources, personal and company-wide alike, and the files in your projects. How large that pool is depends on your licenses:

| License | Storage | | ------------ | ------------------------------------------ | | Business | 1 GB per license, shared in the workspace | | Max | 10 GB per license, shared in the workspace |

The figure is calculated per license, but the space belongs to everyone together: 20 Business licenses add up to 20 GB for the whole company, no matter who uses how much of it.

How much is already taken is shown to administrators at the top of Settings → Data sources. The bar shows separately how much goes to connected systems and how much to projects, with every source and every project listed underneath, largest first. That makes it obvious where there is something to clear out.

When storage is full, the two paths behave differently. Anything already indexed stays searchable either way. New files from a connected data source wait for space and are picked up automatically once there is some. An upload into a project, on the other hand, is refused right away, because somebody is standing there and would rather hear at once that it did not work.

You free up space by deleting data sources or projects you no longer need.

Buying more storage. If that is not enough, you can expand at any time. Administrators find the “Add storage” button next to the usage display; it suggests right away how much is missing to cover the files that are waiting. The space is available within seconds, the pro-rated amount for the current billing period is charged immediately to your stored payment method, and waiting files are picked up automatically. The same add-on storage can also be adjusted when buying and managing licenses, where it can be reduced again as well.

What are data sources?

Data sources connect external file storage with 9brains. Documents from these stores are indexed automatically and become available to the AI as searchable knowledge.

This approach is also known as RAG, short for Retrieval-Augmented Generation. The idea: before the AI formulates an answer, it first searches the connected documents for relevant information. That way, answers are not only based on the general language model but on the actual content from your files, with source references.

Without data sources the AI only uses knowledge from manually maintained knowledge bases.

With data sources it can additionally:

Search documents in the connected stores
Provide answers with source references to the original files
Automatically detect and index new and changed files

Which stores can I connect?

OneDrive. The personal Microsoft 365 drive, every user connects their own.

SharePoint. Document libraries from Microsoft 365 as a shared source for the whole workspace, set up by an administrator.

File server. Windows file servers, NAS systems and SMB shares from your own company network, connected through the on-premises connector. Optionally with the Windows file permissions carried over, so each user only sees in the chat what they are allowed to see on the drive.

Microsoft 365 mailbox. A mailbox including its attachments as a searchable source.

Two types of data sources

Personal data sources

Personal data sources are visible only to the respective user. Other users and administrators cannot view the file listings.

Example: An employee connects their personal OneDrive. Only they can search the files from it in the chat.

Workspace data sources

Workspace data sources are available to all users in the workspace. They are set up by administrators and are well suited for shared files.

Example: An administrator connects a SharePoint Document Library with company documents. All employees can search these documents in the chat.

Outlook: more storage types in planning

In addition, S3-compatible storage (e.g. Amazon S3, MinIO) will soon be supported as a native data source with automatic indexing.

Overview: What can I do?

| Action | Who can do it? | | ------------------------------- | -------------- | | Connect a personal data source | All users | | Search files in the chat | All users | | Connect a workspace data source | Administrators | | Trigger a manual sync | Owner / Admin | | Delete a data source | Owner / Admin |

The data source management is accessible via the Settings:

Settings, Data sources

The page shows three areas:

My data sources: personal data sources of the current user
Workspace data sources: available to all users in the workspace
Personal data sources (admins only), data sources of other users

Each data source shows:

Name and type of the source
Status: active, sync running, error
Last sync: time of the last successful sync
File count: how many files are indexed

Supported file types

Data sources index all file types that 9brains can process:

| Category | File types | | --------- | ------------------------------ | | Documents | PDF, DOCX, PPTX, XLSX | | Text | TXT, MD, CSV, HTML | | Images | PNG, JPG, JPEG, TIFF, BMP, GIF |

Unsupported file types (e.g. videos, ZIP archives) are skipped automatically. How many files that affects is shown on the data source’s status card after a sync. On file servers that have grown over the years these are mostly the pre-2007 Office formats (.doc, .xls, .ppt): they are not the documents listed above and are not read. Saving them once in a current format makes them searchable at the next sync.

Synchronization

Data sources are synchronized automatically:

Initial indexing: When the data source is created, all files are indexed
Real-time updates: OneDrive and SharePoint report changes on their own, so changes are typically detected and processed immediately
Regular reconciliation: In addition, a full reconciliation runs every 60 minutes. Stores that do not report changes themselves are covered exclusively by it, so new files there become searchable within an hour
Change detection: Only new, changed, or deleted files are processed, no re-indexing of unchanged files

You can also trigger the sync manually: click on the data source and choose “Start sync”.

When the connection stops working for good

If the sync keeps failing, 9brains pauses it and sets the data source to Error. This happens right away when the sign-in is no longer valid (for example because the Microsoft account has to be reconnected, or a permission was revoked), and after several unsuccessful attempts when the storage stays unreachable.

Files that are already indexed remain searchable, only new ones stop arriving. Once the problem is fixed, start the sync once by hand via “Start sync”; it then continues automatically as before.

A temporary glitch, such as a file server that is briefly unavailable, does not trigger this pause: 9brains keeps trying on its own.

Frequently asked questions

Which license do I need?

Data sources are available starting with the Business license. The feature is not accessible with the Pro or Starter license.

What happens when storage is full?

See Storage: what is already there stays searchable, new files from a data source wait for space, and a project upload is refused.

Can I connect multiple OneDrive accounts?

No, one personal OneDrive can be connected per user. For shared files, SharePoint data sources are the right fit.

Can I connect multiple SharePoint sites?

Yes. Administrators can connect any number of SharePoint Document Libraries as separate workspace data sources, as long as the storage quota is sufficient.

Are deleted files removed from the index immediately?

Deleted files are typically detected immediately and removed from the index. At the latest, they are removed during the next regular reconciliation (every 60 minutes).

Who can see my personal data sources?

No one but you. In the overview, administrators only see that a personal data source exists, the file listings are not visible for privacy reasons.