There are four ways to get files into docuflow-ai. All of them feed the same pipeline, so a document behaves the same no matter how it arrived.
| Channel | Best for | Set up by |
|---|---|---|
| Browser upload | Ad-hoc files and first tests | Any user with the Document upload page |
| Capture email | Suppliers or colleagues who send attachments | Nothing; every organization gets an address |
| REST API | Line-of-business systems and integrations | A developer, with your API key and secret |
| Link collections | Scheduled collection from web portals and file locations | A platform administrator |
Upload from the browser#
- Open Document upload from the left menu.
- Add files by dragging them onto the drop area or with Select files. Each file is listed with its size and status.
- Click Upload. A floating progress panel tracks the transfer, so you can move to other pages while it runs. Use Clear to empty the list.
| Rule | Detail |
|---|---|
| Maximum file size | 200 MB per file. |
| Accepted extensions | Only active content types. Other files are rejected with Not supported. |
| Duplicates | A file with the same content (CRC32) and size as an existing document is marked Duplicate and handled by DefaultDuplicateBehaviour: skip it (default) or overwrite the existing file. |
| Ownership | The document is recorded under your user; API and email uploads are recorded under the system user. |
Capture email#
Every organization has a private capture address in the form <your API key>@<capture domain>.
You can copy it from Admin → Organization Management → Initial Setup once it has been provisioned.
- Send or forward an email with attachments to the address. Every file attachment with a supported extension becomes a document in your organization.
- The mailbox is checked on a schedule by the Capture mail task. Processed messages are marked as read and later removed by Delete handled mail.
- Messages without attachments are ignored; the email body itself is not stored.
- Mail that doesn't match any organization is quarantined and logged in Error Logs.
REST API#
Systems can push files with POST /api/documents (Base64 content) or register files that already sit in the
organization's storage folder with POST /api/documents/record and /record/bulk (up to 500 at a time).
Afterwards they can follow the processing with GET /api/documents/{id}/status and read the documents back with
GET /api/documents. See the API reference for request formats and code samples.
Link collections#
A link collection is a list of sources (for example supplier portals) that the DocuRunner batch runner visits on a schedule to collect documents on your behalf. Collected files are registered through the API and enter the pipeline like any other upload.
- Create a schedule under Admin → Link collections → Link Collection Schedules. A schedule has a name, a cron expression and an enabled flag.
- Add links under Link Collections: name, description, URL, file path, link type, optional credentials, expiration date and schedule.
- Bulk edit with Export template and Import data, then push the definitions with Export to DocuRunner. Row edits are synchronized automatically.
What happens after upload#
Uploading only stores the file. Splitting, versioning, text extraction, indexing and field extraction happen in the background; follow progress in the Status column of the Documents page. See Core concepts & pipeline.