- Documentor
- Groups and access
Everyone sees what they should. Everyone manages what is theirs.
Archives belong to user groups. The group decides who administers an archive, where its files are stored and which AI models are used. Roles are assigned to individual people as well as whole groups, and everyone can see on their profile where each role comes from.
Application, group, archive.
Three levels, each with its own responsibility. The Supervisor looks after the whole installation, the group administrator their group and its archives, the Archive administrator the archive’s settings and permissions. Nobody has to hold more rights than they need.
Supervisor
Approves new users, creates groups, appoints their administrators and decides on the configuration. Sees all archives, the processing queue, the log and usage.
- Role from a group in the application or from Entra ID
- Approves proposals from group administrators
- The DEFAULT group for everything not set up elsewhere
Group administrator
Adds and removes members, is the Archive administrator of all the group’s archives and may create new ones in it. Sees the members’ usage and proposes changes to storage and models.
- Members get the group’s roles all at once
- Its own S3 storage and AI models
- Default notifications for the members
Archive administrator
Decides what happens during processing, writes instructions for the AI, labels and code rules, and assigns the roles Reader, Processor and Archive administrator to people and groups.
- Its own documents, index and permissions
- Optional processing phases
- External AI only with the archive’s consent
Who can do what.
Three roles apply to each archive separately, two to the whole application. A person can have a role directly, through a user group in the application or through a group in Entra ID; an assigned role takes effect once the page is reloaded.
| Action | Reader | Processor | Archive administrator | Archive creator | Supervisor |
|---|---|---|---|---|---|
| search, read, ask the AI | yes | yes | yes | in their own archives | yes |
| upload and delete documents, process again | – | yes | yes | in their own archives | yes |
| sort incoming documents into the archive | – | yes | yes | in their own archives | yes |
| change the archive’s settings and permissions | – | – | yes | in their own archives | yes |
| create archives | – | – | – | yes | yes |
| approve users, manage groups, the queue, the log and usage | – | – | – | – | yes |
Administration without the Supervisor
The Supervisor appoints a group administrator from among the group’s members. The group administrator is the Archive administrator of all the group’s archives, may create new ones in it, adds and removes members and sees their usage. A group that grants the Supervisor role can be edited only by a Supervisor.
No way around access
A Supervisor cannot remove their own role or block themselves; they have to ask another Supervisor. A blocked user has no access, not even through a group. A sorter who files a document into an archive does not gain access to that archive by doing so.
Access for a whole team in one step.
When several people need the same access, the role goes to a group. A new member gets the group’s roles straight away, a removed member loses them. Every user is in at least one group.
Google accounts too
A group kept in Documentor gives its members roles in archives and, optionally, a role in the application. It works for Microsoft and Google accounts as well as accounts from other organisations.
Your organisation’s groups
A role in an archive can also be given to an Entra ID group by its Object ID. It applies to your organisation’s Microsoft accounts, and you manage membership where you always have.
Members of several groups
They choose a default group in the menu under their name. It is pre-filled when they create an archive, and the application uses it to choose the AI models for their chat and search.
Nothing is left without a group
The application’s default group: it owns the archives from before groups existed, and anyone who would otherwise be in no group moves into it. Whatever a group does not set up is taken from DEFAULT.

Separate storage and AI models for each group.
Each group can have its own file servers and its own AI models. An archive only chooses from them and never gets the addresses or keys. Useful when each company or department should store its files elsewhere or pay for AI separately.
- S3 storageany S3-compatible storage (SeaweedFS in our case) for originals, page images and PDFs with text – on your premises too; the group’s default storage is pre-filled when an archive is created
- Moving an archivean archive can be moved to other storage while in use; during the move it can still be searched and read
- AI modelslanguage model, embedding and reranker; for each, whether it runs at a third party and whether it understands images
- Default modelsfor the group’s archives without a choice of their own, and for the chat and research of its members
- Secretspasswords and API keys are stored encrypted and are never shown again once saved
- Supervisor
- changes the configuration immediately
- Group administrator
- proposes a change for approval
- Until approved
- the current settings apply, and the item shows change pending
- Before approval
- the Supervisor tests the connection to the proposed server
- Rejection
- always with a reason, which the group administrator sees with the change
- Conflicts
- if the connection has changed in the meantime, the proposal is outdated and nothing is applied
No passwords of its own, with approval.
Documentor has no passwords of its own. Anyone who signs in for the first time registers themselves and waits until the Supervisor approves them – or until an Archive administrator assigns them a role.
Microsoft or Google account
An organisation or personal Microsoft account, or a Google account. Which buttons are offered depends on the installation; the Admin role from Entra ID makes a user a Supervisor.
Archives, roles and groups in one go
When approving a user, the Supervisor picks archives with a role, and groups. A user can be rejected or later blocked; their roles are kept in case they are approved again.
Where each of my roles comes from
On their profile everyone sees their groups, and their archives with the role and where it comes from: directly, through a group, through an Entra ID group, or as a group administrator. That is also where they set their default group and notifications.
You know what goes out and what it costs.
A model at a third party gets documents only from an archive that has explicitly allowed it. How much was searched, asked of the AI and processed is shown in Usage – separately for external and internal models.
- External AI
- Archive consentthe Allow external model switch in the archive settings; a model on your own server needs no permission
- What is sentpage crops, text excerpts for answers and the text of documents for annotations – only from an archive that allows it
- Without permissionsteps with an external model are skipped, search keeps working
- Off-peaksteps with external AI can wait for the cheaper tariff, the rest of the processing carries on
- Usage
- By usersearches, AI questions, research and tokens, broken down by model
- By groupthe total for its members; a group administrator sees their own members’ usage directly
- By archivedocuments, pages, storage, index and processing tokens – they count towards the archive, not whoever uploaded the documents
- Exporttables to Excel as a basis for recharging costs
You hear about what concerns you.
For each kind, a person chooses immediately (notifications from a few minutes in one e-mail), daily digest at 7:00 or off. A group administrator sets the default for the members; a personal choice takes precedence.
Requests and proposals
The outcome of a request for access or of a proposed configuration change; for the Supervisor, new users and changes awaiting approval.
What went wrong
Documents in the archive that ended with an error, and a failed storage move – for Processors, administrators and the Supervisor.
What has been added
Newly processed documents in the archives you have access to. Off by default.
What is waiting
In the daily digest: how many incoming documents are waiting to be filed. For sorters, when document intake is switched on.
We will show you how it works on your own documents.
Send us a few PDFs from your archive – old scans are fine. In 30 minutes you will see how searching them works and how the chat answers.
Already a customer? Sign in