Documents

Documents in SeaSquad are organised per vessel, processed automatically on upload, and made searchable — both directly and through the AI assistant — without any manual tagging.

Folders per vessel

Each ship has its own set of document folders. Folders can be used to group documents by type or purpose — manuals, survey reports, correspondence, and so on — matching how your team already organises vessel paperwork. You can browse a vessel's folders in a grid or list view, and navigate into a folder to see the documents inside it.

Uploading documents

You can upload a single file or select multiple files at once for a bulk upload into a folder. Files can be dragged and dropped or selected from your device.

Supported formats

• PDF and images (PNG, JPEG, GIF, WebP) get the fullest treatment — these go through vision-based text and table extraction, which handles scanned pages, photographs of documents, and typed originals equally well.
• Office documents — Word (.doc, .docx) and Excel (.xls, .xlsx) — and plain text files are also supported and processed for search and metadata.

There's a per-file size limit of 50 MB.

What happens after upload

Once a document is uploaded, it moves through an automatic processing pipeline:

1. Text extraction. The document's content — including tables — is extracted using vision-language models, which read the page as an image rather than relying on the file having clean embedded text. This is what makes scans and photographed pages work as well as native PDFs.
2. Content chunking. Extracted text is split into chunks so the AI assistant can retrieve the specific, relevant part of a long document rather than the whole file.
3. Metadata extraction. Key details — document type, dates, identifiers, and other structured fields depending on the document — are pulled out automatically.
4. Indexing. Once processed, the document's content is searchable in the portal and available to the AI assistant.

This typically takes a few minutes per document, depending on length and how busy the processing queue is.

Processing states

Each document shows its current state so you can tell when it's ready:

A visual indicator on each document shows which step it's currently on, so you can see progress rather than just a spinner.

If a document fails to process, you can retry the upload; if it continues to fail, check that the file isn't corrupted or password-protected, and contact your administrator if the problem persists.

Replacing files

If you need to upload a newer version of a document — a corrected report, a reissued manual — you can replace the file on an existing document entry rather than creating a duplicate. The replacement goes through the same processing pipeline, and the document's extracted content and metadata are refreshed accordingly.