Skip to content
Deltacraft

Document capture

Live

Cerno

Scan it once. Never type it again.

Cerno turns a stack of paper into filed, searchable, indexed records without an indexing clerk retyping it. Pages arrive from a scanner, a watched folder, an upload or the API, get read, classified and extracted, and come out as an encrypted searchable PDF/A in the right place in the filing plan.

The reader is the point. Because OCR now reads handwriting as well as print, the operator confirms what the machine read instead of keying it in, and only the documents that fall below the confidence policy for their type reach a person at all. Cerno runs on your own servers or as a desktop install, so the documents never leave the institution that owns them.

At a glance

OCR engine
PP-OCRv5OCR engine
Archive format
PDF/A-2bArchive format
Encryption at rest
AES-256Encryption at rest
Scanner protocols
3Scanner protocols
Export destinations
8Export destinations
Interface languages
4Interface languages

Connects to

  • VaultIQ
  • SharePoint
  • Dynamics 365 Business Central
  • Scanners (TWAIN, SANE, eSCL)
  • FTP, webhook and email

From paper to system of record

The chain this product holds.

Each link is an artefact with an owner and a date. Cerno will not let you create one without the one before it, which is what makes the trail defensible after the fact.

  1. 01

    Capture

    Scanner, operator, batch ticket

  2. 02

    Separate

    Barcode separator, blank pages removed

  3. 03

    Read

    OCR text, per-word confidence

  4. 04

    Classify

    Document type, evidence that decided it

  5. 05

    Extract

    Field, value, position on the page

  6. 06

    Corroborate

    Matched record, source system

  7. 07

    Verify

    Who confirmed, what they changed

  8. 08

    Deliver

    Archive path, destination, receipt

Capabilities

What Cerno does.

Capture

Get the paper in without a keyboard.

Scanner bridge
TWAIN, SANE and eSCL, so the scanners already on the desks are the supported ones rather than a model you have to buy.
Batch separation
Barcode separator sheets split a stack into documents, and blank pages are dropped on the way past.
Hot folders and API
Watched folders, uploads and an API for everything that was already digital, on the same pipeline as the scanned paper.
Duplicate detection
File hash, perceptual hash, content similarity and field matching catch the same page scanned twice, in two batches, weeks apart.

Read

One reader, and it reads handwriting.

PP-OCRv5 recognition
A single reader trained on handwriting as well as print, reporting per-word confidence rather than a yes or no.
Classification ladder
Layout fingerprint, then similarity to approved examples, then keyword hints. A model is called only on cold start.
Field extraction
Learned zones, label anchors and format rules pull the values out, and each one keeps the position it came from on the page.
System-of-record lookup
A match in Business Central or another source both verifies the key field and populates the rest at high confidence.

File

An archive that survives an audit.

Verification as spot-check
A confidence policy per document type decides what a person sees. The rest files itself, and the exceptions are the queue.
Searchable PDF/A
A text layer over the original image, written to PDF/A-2b so the archive is still readable when the software is not.
Encrypted archive
AES-256-GCM at rest per tenant, filed against a filing plan rather than dropped into a folder someone has to remember.
Delivery with receipts
VaultIQ, SharePoint, another EDMS, a network path, FTP, webhook, email or a watched folder, with a retry queue for when a destination is down.

Built for

Who runs this.

  • SACCOs and member-based institutions
  • Banks and insurers with paper files
  • Records, registry and archive teams
  • Shared services running a scanning bureau

Cerno for your team

See Cerno against your own process.

We will walk it through with your data and your workflow rather than a demo dataset, so you can tell whether it fits before anyone signs anything.

Nairobi office

10th Floor, Applewood Adams, Ngong Road
Nairobi
Kenya
Hours
Mon–Fri, 08:30–17:30 EAT