ABBYY adds explicit PDF text-versus-OCR processing controls to Vantage
ABBYY introduces PDF processing controls in Vantage, allowing enterprise document workflows to choose between embedded text and OCR, with Technology Core 3 required.
The new setting lets document teams choose automatic PDF handling, reuse trusted embedded text or force OCR when existing text layers are unreliable. Technology Core 3 is required.
ABBYY detailed its new Vantage PDF Processing Mode on 13 October 2026, giving teams explicit choices over whether the intelligent document processing platform reads a PDF’s embedded text or performs optical character recognition (OCR). The setting is intended for workflows handling mixtures of digital documents, scanned pages and files with unreliable text layers.
In its official product announcement, ABBYY described three modes: Default, Use Text Layer Only and Use OCR Only. The choice affects how Vantage reads PDF content before downstream extraction, rather than introducing a separate document processing engine or a new Vantage platform.
Three modes control how PDF content is read
Default retains Vantage’s automatic behaviour. It uses an embedded PDF text layer when present and supplements it with OCR when necessary, making it suitable for collections containing both digitally generated and scanned documents.
Use Text Layer Only prioritises the existing text embedded in a PDF instead of performing full-page OCR. ABBYY says this can suit trusted files, such as invoices generated directly by enterprise software. Despite its name, the option falls back to OCR if a document has no text layer; it does not leave scanned PDFs unread.
Use OCR Only disregards the embedded text layer and runs OCR across every page. The option is designed for documents whose text layers are corrupted or unreliable, including files produced by older scanning and recognition systems.
Corrupted text layers are the problem ABBYY is targeting
Some PDF files display readable-looking pages while containing incorrect machine-readable characters underneath. Downstream extraction may use that faulty embedded text, introducing errors even when the visible document appears normal. ABBYY’s new setting allows a workflow designer to make the OCR-versus-text-layer decision explicitly rather than depending entirely on automatic selection.
The distinction matters in document-heavy processes such as invoice handling or records intake, where files can come from unrelated software and scanning systems. Choosing OCR for problematic sources may avoid reusing a damaged text layer, while trusting a clean text layer can avoid unnecessary recognition work.
ABBYY presents these outcomes as expected benefits of more predictable PDF handling. Its October explanation does not provide independent accuracy comparisons or quantified processing-speed results for the three modes.
Configuration is limited to Technology Core 3
ABBYY says PDF Processing Mode is available in Vantage Skill Designer under Image Processing on the General tab, either within an OCR Skill or an OCR activity inside a Process skill. The controls are supported only for skills using Technology Core 3; earlier Technology Core versions do not have the setting.
For organisations using older skills, that requirement makes the technology core version a prerequisite to changing PDF processing behaviour. ABBYY’s 13 October explanation does not specify a separate customer rollout timetable or independently verified performance benchmarks for the update.