Paperless-ngx
Watches a folder or email inbox, runs OCR on incoming scans, and suggests tags, correspondents and document types. It stores the original file alongside a searchable PDF version, which is worth allowing for when sizing disk.
Document archive server in Kenya
Turn the box files into a searchable archive
Invoices, contracts, court bundles and payroll records pile up in box files that flood, fade and take hours to search. A storage server gives your firm space to scan everything once, run OCR, and find any page by typing a name, PIN or case number.
300 thousand pages
We suggest
Storage S1
KES 2,000/mo
About 700,000 pages at roughly 1 MB each is around 700 GB, within the 1 TB disk. Note OCR on a single core is slow for large backlogs.
Order Storage S1Paper records are a risk you keep paying rent for
An accounting firm with forty clients can easily hold several hundred box files: bank statements, ETR receipts, payroll sheets and signed financial statements. A law office has client files, title documents and correspondence going back decades. Both pay for shelf space and lose hours every month looking for one sheet of paper.
Scanning is the easy part. The hard part is putting the scans somewhere that will still be there and still be searchable in five or ten years. A desktop folder full of files called scan0043.pdf is not an archive. A document management system such as Paperless-ngx or Mayan EDMS reads the text with OCR, tags each file by client, year and type, and lets staff search across everything from a browser.
Those systems need a server that stays on, with enough disk for originals plus the searchable copies the software produces. A storage plan provides that capacity, with root access to install the software, and room to grow as you work through the backlog. Keep the paper you are legally required to keep until your adviser says otherwise; the server makes it findable and protects you against fire, water and lost files.
The problem
Problems a document archive solves
Hours lost to searching
A KRA query, an audit request or a client dispute asks for one invoice from three years ago. Staff pull box files from a storeroom and page through them. With OCR search the same request takes a minute.
Fire, water and pests
A leaking roof during the long rains, a small office fire or termites can destroy records you are required to keep. Scanned copies held off-site mean the loss of paper does not mean the loss of the information.
Too many people with access
Open shelves let anyone read payroll, HR files and client agreements. A document system gives each user access only to the clients or departments they work on, and keeps a record of who opened what.
Scans nobody can find
Firms that already scan often end up with thousands of unnamed PDFs across laptops and email. Without OCR and consistent tags, the archive exists but is barely more useful than the box files it replaced.
What runs on it
Document software you can install
These open-source tools are widely used for business archives. With full root access you install and configure them yourself, usually with Docker.
Mayan EDMS
A fuller document management system with workflows, metadata fields, version history and detailed permissions. A better fit than Paperless-ngx for a law firm that needs approval steps or per-matter access rules.
OCRmyPDF
A command-line tool that adds a text layer to scanned PDFs. Useful for batch-converting an existing folder of old scans before importing them, and it is the OCR engine Paperless-ngx relies on.
Nextcloud
Gives staff a familiar file-and-folder view with desktop sync, and lets you share a specific folder with an external auditor for a limited time instead of emailing attachments.
BorgBackup
Deduplicating, encrypted backups of the document database and media folder. Use it to keep dated copies on the server and to send a second copy elsewhere, so a bad import or deletion can be reversed.
Sizing
Sizing for scanned pages
Set the slider to the number of pages you plan to keep. We assume about 1 MB per page stored, covering a 300 dpi scan, the searchable copy and thumbnails, with roughly 25% of the disk left free.
- 10–700 thousand pages Storage S1 About 700,000 pages at roughly 1 MB each is around 700 GB, within the 1 TB disk. Note OCR on a single core is slow for large backlogs.
- 710–3,000 thousand pages Storage S2 Up to three million pages is about 3 TB of 4 TB. The two cores and 8 GB RAM also make Paperless-ngx OCR noticeably quicker on bulk imports.
- 3,010–6,000 thousand pages Storage S3 Six million pages is roughly 6 TB on an 8 TB disk. Four cores let a multi-branch firm run OCR on several scanner queues at once.
- 6,010–9,000 thousand pages Storage S4 Around nine million pages uses about 9 TB of 12 TB, leaving space for database dumps and Borg snapshots of the archive itself.
- 9,010+ thousand pages Storage S5 Up to about twelve million pages fits the 16 TB disk with headroom. Archives larger than that are usually split by department or year.
This is an estimate: colour scans and photographs of documents take several times more space per page than black-and-white text, and the scanner's resolution and compression settings matter most.
Storage S1
KES 2,000/mo
- 1 core
- 2 GB RAM
- 1 TB SATA storage
- 2 TB transfer
Storage S2
KES 5,000/mo
- 2 cores
- 8 GB RAM
- 4 TB SATA storage
- 8 TB transfer
Storage S3
KES 7,000/mo
- 4 cores
- 16 GB RAM
- 8 TB SATA storage
- 16 TB transfer
Storage S4
KES 15,000/mo
- 6 cores
- 24 GB RAM
- 12 TB SATA storage
- 24 TB transfer
Storage S5
KES 20,000/mo
- 8 cores
- 32 GB RAM
- 16 TB SATA storage
- 32 TB transfer
Before you go live
Setting up a document archive
0 of 8 done. Tick them off as you go.
Built for Kenya
Record keeping in Kenya
Tax records have a retention period
The Tax Procedures Act, 2015 generally requires taxpayers to keep records for five years, and longer in some situations. Digital copies help you meet that period, but confirm with your tax adviser which originals you must still keep on paper.
Client files hold personal data
ID copies, KRA PINs, payslips and contracts are personal data under the Data Protection Act, 2019. Keep access limited to staff who need it, encrypt backups, and delete records once the reason for keeping them has passed.
Scan during the day, upload at night
Office internet is often shared with calls and client work. Let scans queue locally and sync to the server overnight, and check your upload speed so you know how quickly a backlog of box files will clear.
Questions
Document archive: your questions, answered
Are scanned copies acceptable to KRA instead of paper?
Whether a scanned copy can stand in for a paper original depends on the type of record and on current KRA guidance. Treat the archive as your searchable, protected copy, and confirm with your tax adviser before destroying any originals that may be needed for an audit.
What is the difference between Paperless-ngx and Mayan EDMS?
Paperless-ngx is simpler: scan, OCR, tag and search, with very little setup. Mayan EDMS adds workflows, custom metadata, document versions and finer permissions. A small accounting practice is usually well served by Paperless-ngx, while a larger law firm may prefer Mayan for matter-based access control.
Will OCR read handwritten notes or poor-quality copies?
OCR handles clean printed text well, including most invoices and typed letters. It struggles with handwriting, faded thermal receipts and skewed photocopies. Tag those documents carefully by client and date so they can still be found, and scan thermal receipts promptly before they fade.
Which plan suits a small accounting firm?
Most small firms start on Storage S1 or S2. Disk space is rarely the limit at first; the processor matters more because OCR is demanding. If you are importing a large backlog, the extra cores and memory in Storage S2 make the first months of scanning much faster.
Can our external auditors see selected documents?
Yes. Create a separate account in your document system limited to the client and year being audited, or share a Nextcloud folder with a password and expiry date. Remove the access when the audit closes, so nobody keeps a standing login to your archive.
Do you scan or manage the documents for us?
No. The plan is a storage server you control. You or your IT provider install the archive software and handle scanning. We provide the server, disk capacity and network, and you keep full control of the records and who can see them.
Also on Storage Hosting
Backup storage for what you need to keep
Ready to host your document archive?
Plans from KES 2,000 a month, paid by M-Pesa or card.
Size your server