Self-hosted intelligent document processing

Capture every page.
Pay for none of them.

KognitCapture is a multi-tenant, self-hosted capture platform that replaces ABBYY FlexiCapture, ABBYY Vantage, Kofax/Tungsten, ChronoScan, Rossum and Ephesoft — with no per-page meter, no station licences and no module unlocks. Your documents stay on your hardware. Your bill stops moving when your volume grows.

Runs on Linux and PostgreSQL. Browser-only operator hubs — nothing installed on a desk. Tesseract, local GGUF models or thirteen remote LLM providers, switchable per project.

Scanned customer purchase order with recognised field zones drawn over it
customer_name
po_number
order_date
order_lines
vat_amount
customer_nameDelta Components NV99%
po_numberPO-2026-0441798%
order_date2026-08-1794%
order_lines5 rows · 4 columns92%
vat_amount452,4068%
€0per page, per document, per transaction
83capabilities benchmarked against nine capture platforms
13 + 13import sources and export targets, first-class
129OCR language packs, local by default
109MCP tools for agents, behind OAuth 2.1 and scopes
1 installserves every tenant you onboard

Switching from

Pick the platform you are replacing.

Each page is a straight read: where that product genuinely beats us, where we beat it, and what the move actually costs. We publish both columns, because a comparison that only lists strengths is worth nothing to you in an evaluation.

Incumbent · on-prem

ABBYY FlexiCapture

The deep on-premises capture platform we were scoped to replace: Windows/IIS/MS SQL, FlexiLayout templates, thick-client stations, invoice packages.

Read the comparison →

Cloud IDP

ABBYY Vantage

ABBYY's strategic bet: skills-based, pre-trained, low-code, container-deployed — fast to value, priced per transaction, and the models are ABBYY's.

Read the comparison →

Conversion server

FineReader Server

High-throughput page conversion rather than data capture: scan-to-searchable-PDF/A at volume with light indexing.

Read the comparison →

Incumbent · platform

Kofax / Tungsten

Capture for ingestion plus TotalAgility for orchestration, RPA and GenAI. Widest footprint in the market, and the price list to match.

Read the comparison →

Template-free IDP

Rossum

Cloud-only transactional document AI, part of Coupa. Its Aurora model reads unseen layouts without templates and learns from every correction.

Read the comparison →

SMB capture

ChronoScan

Windows desktop/server capture: zonal and dual-engine OCR zones, barcode, hot folders, bidirectional ODBC, AI extraction through your own provider key. Perpetual licence from ≈ €245.

Read the comparison →

On-prem capture

Ephesoft Transact

Java capture platform now inside Tungsten Automation, built around batch classes and automatic document classification — architecturally the closest peer in this market.

Read the comparison →

Cloud OCR API

Amazon Textract

Five APIs returning text, forms, tables, queries and signatures as JSON, priced per page. No workflow, no templates, no operator UI — you build everything around it.

Read the comparison →

AP specialist

ReadSoft

Invoice capture welded to the ERP: ReadSoft Invoices / Online plus Process Director inside SAP, matching live master data and posting in-system.

Read the comparison →

Why buyers move

Four arguments the incumbents cannot answer at once.

Unit economics

Your bill stops scaling with your volume

Metered platforms charge roughly $0.02–0.10 a page (public list bands). At two million pages that is a page bill on top of a platform fee. Here recognition runs on your own CPU with Tesseract or a local model: the marginal cost of the next page is the electricity to read it. Use a remote LLM if you want to — tokens are logged and costed per tenant, and you decide which projects may call out at all.

Sovereignty

Documents never leave your estate

Self-hosted on Linux and PostgreSQL, air-gap capable, with local GGUF models for both text and vision. Tenant → organisation → project isolation reaches down to the storage prefix, and every agent, API key and operator action lands in an audit channel. There is no licence server to phone home to, because there is no licence gate.

Operations

Nothing installed on an operator's desk

Scanning, processing and verification hubs are browser pages. Network scanners are driven server-side over eSCL, so a scanning desk needs a browser and nothing else — no thick client, no TWAIN driver rollout, no per-station licence. The hubs also embed into your own application behind an API key, which is how software vendors ship capture without building it.

Multi-tenancy

One install, every client you onboard

FlexiCapture, ReadSoft and ChronoScan expect one deployment per customer. This platform was built tenanted: a service provider runs twenty clients on one cluster, with per-tenant storage, roles, retention, LLM budgets and audit. That is the single strongest number in any business case a BPO or MSP puts together.

Timing

ABBYY is moving FlexiCapture accounts to Vantage. That is a rebuild, not a transfer.

There is no converter that turns a FlexiCapture document definition into a Vantage skill — practitioners describe the path as inventory, map, rebuild, re-integrate, run in parallel, and ABBYY deployments are commonly quoted at eight to twelve weeks. If you have to rebuild anyway, the honest question is no longer “replace something that works” but “rebuild there, or rebuild here”.

Read the migration brief →

Where we say no

Four situations where we will tell you to buy something else.

  • SAP-native accounts payable posted inside the ERP — ReadSoft owns that; we feed it.
  • A TWAIN or ISIS-only production scanner estate with no eSCL support.
  • Procurement that must shortlist from the IDP Magic Quadrant.
  • Buyers who need SOC 2 or ISO 27001 held by the software vendor, not the operator.

We would rather lose the deal in week one than in month six.

See it working

The demo is the product's own interface, not a slide.

Project configuration, the scanning hub with its assembly board, and the verification hub with zone overlays, confidence bars and validation errors — rendered from the application's own stylesheets and markup.

Questions we get in week one

Straight answers, including the awkward ones.

Is open-source OCR really good enough to replace ABBYY?
On clean 300 DPI print, Tesseract 5 is competitive. On degraded scans, dense forms and non-Latin scripts, ABBYY's engine still reads measurably better — we will not pretend otherwise. The answer for those documents is the hybrid path: Tesseract first, and low-confidence words or whole pages re-read by a vision model, local or remote, at your choice. The only way to settle it is a bake-off on your own documents — a paid engagement at our consulting rate, which is what we propose in every evaluation.
What does it actually cost?
A platform licence for the deployment, not a meter on your documents. Above roughly 100,000–150,000 pages a year, self-hosting is decisively cheaper than a metered IDP service and the gap widens every year. Below that, a metered platform is often genuinely cheaper than running infrastructure — we will say so. Run the numbers on the pricing page.
Do we have to send documents to an AI provider?
No. Recognition can be entirely local: Tesseract for text, a local GGUF model for vision through a managed llama-server. Thirteen remote providers are supported for teams that want them, per project, with tokens and cost logged per tenant. An air-gapped install is a supported deployment, not a workaround.
Can we keep our scanners?
If they speak eSCL (AirScan/Mopria) — which covers most networked MFPs and modern departmental scanners — yes, and the server drives them directly, so operators install nothing. Older production scanners that are TWAIN or ISIS only cannot be driven at all. That is a hardware fact, and we ask what is on your floor before the demo rather than after it.
What happens to our existing templates?
Field sets, validation rules and export mappings are rebuilt here — the same work any migration between capture platforms requires, including ABBYY's own path from FlexiCapture to Vantage. What shortens it: six industry blueprints that provision a project, workflow, routing, upload endpoint, export goal and a full field set in one action, and model-assisted template authoring that drafts a template from one sample page. Blueprints ship pre-authored field sets, not pre-trained models — see the migration brief for what that means in weeks.
Is the bake-off free?
No. The 15-minute scanner and volume call is free, and so is the demo. The bake-off is paid work at our standard consulting rate, scoped and quoted before we start: someone reads your documents, authors templates, runs Tesseract, the hybrid path and the model path over them, and writes up field-level accuracy with the cost per thousand pages of each mode. It runs one to two consulting days — two is the ceiling, never a week. Pilots are quoted the same way, separately from the platform licence. You keep the report whatever it concludes — including when it concludes that a competitor reads your paper better.
Who is behind it, and what if you disappear?
KognitCapture is self-hosted and source-available to customers: there is no licence server that can switch your installation off, and an escrow or source arrangement is part of the standard enterprise conversation rather than an exception to it.