Upload your first documents

A Muninnbase tenant is only as useful as the documents inside it. An empty knowledge base is not a knowledge base. The fastest way to feel the product working is to upload five or ten of your most-referenced files and ask it a few questions you already know the answers to.

What to upload first

Start with the documents your team asks about most often. For a typical small business, that's something like:

• Employee handbook
• PTO and time-off policy
• Expense reimbursement SOP
• IT onboarding guide
• Dress code or uniform policy
• Benefits summary

Five to ten files in this category is enough to start. You can add the rest later; the goal of this first batch is to confirm the system works against questions you can verify yourself.

Supported file types

Muninnbase accepts PDF, DOCX, XLSX, RTF, and TXT. PDFs work best when the text is selectable (not a scan). DOCX files keep their heading structure, which improves how the system chunks them. XLSX files are read row by row with the header row prepended to each chunk, so column labels travel with every cell.

A note about sensitive information

Do not upload anything containing personally identifiable information about your employees. SSNs, bank account numbers, salary records, performance reviews, anything you would not put in a shared drive your whole company can read. Muninnbase is for reference documents that the whole team is allowed to see. It is not a personnel records system.

If your handbook has a section on benefits, that's fine. If your "handbook" is actually a folder of individual employee files, those do not belong here.

How to upload

From /admin/documents, click the Upload button (top right) or drag files directly onto the page. You can select multiple files at once. The upload dialog shows each file's progress and a status badge as it moves through the pipeline.

The page polls automatically, so you can leave the documents tab open and watch the badges change, or come back in a minute.

Status badges and what they mean

Every document carries one of four statuses:

• Pending. The file is uploaded but ingestion has not started yet. Usually a brief state.
• Processing. The parser is running. Text is being extracted, the document is being broken into chunks, embeddings are being computed. Typically a few seconds; longer for large PDFs.
• Indexed. Ready for search. Your team can ask questions about it now.
• Failed. Ingestion broke. See the next section.

When a file fails

The most common cause is a scan-only PDF with no extractable text. The parser cannot read images, so a PDF that is technically just a stack of photographs of pages produces zero chunks and fails. Fix: run the file through an OCR step (Acrobat, ABBYY, or any modern OCR tool) to embed real text, then re-upload.

The other common cause is a corrupted file. The badge tooltip shows the underlying error message. Open the original file in its native application to confirm it works there; if not, get a fresh copy from your source.

A handful of error messages are worth recognizing if you see them in the upload dialog:

• "is too large" means the file is over the 50 MB per-file limit. Split or compress.
• "this file type isn't supported" means we cannot accept the extension. Convert to one of the five supported types.
• "Upload queue is busy" means your tenant has hit the per-tenant pending-uploads cap. Wait a minute for the queue to drain and try again.
• "couldn't be uploaded" is a transient network or server problem. Try again in a moment.

Individual failed files cannot be retried from inside the upload dialog. Close it, fix the file, and re-upload. Files that succeeded on the same batch are unaffected; only the failed ones need to come back through.

How long indexing takes

Most documents finish in a few seconds. A multi-hundred-page PDF can take a minute or two. The documents page is the source of truth for status, so you can leave it open and watch the badges change, or check back later.

The Sync Now button

Sync Now re-enqueues failed or stuck documents back into the ingestion queue so they get another try. Useful right after you've fixed a previously failed file (for example, you OCR'd a scan-only PDF and re-uploaded, but a sibling file still shows Failed from before). Sync Now retries the sibling. For the standard upload flow you do not need it. Most weeks you will never touch it.

What to do next

With documents indexed, the workspace is ready for your team. Invite and manage your team walks through bringing employees in, picking the right role for each, and what to expect when an invite goes out.

If you want to test the system yourself before inviting anyone, head to the chat (/chat) and ask a question whose answer you know is in one of the documents you just uploaded.