1. Presentation

Intelligent OCR Document Management turns Dolibarr into a full Intelligent Document Processing (IDP) platform. It automatically recognises, extracts, classifies and exploits documents — invoices, orders, delivery notes, quotes, contracts, HR, accounting, legal, technical and administrative documents — directly inside Dolibarr, removing the need for a separate OCR product.

It is an integrated alternative to professional solutions such as ABBYY Vantage, UiPath Document Understanding, Rossum, Microsoft AI Builder, Google Document AI, Amazon Textract, Kofax Intelligent Automation and OpenText Intelligent Capture.

2. Architecture

The module follows the native Dolibarr architecture (object-oriented PHP, CommonObject-based classes, standard rights and menu system) and adds a dedicated Intelligent Document Processing pipeline:

·       OCR Engine — printed and handwritten text, tables, stamps, signatures, barcodes and QR codes, automatic rotation and image correction, multilingual recognition.

·       AI Document Understanding — automatic document-type identification and field comprehension.

·       Machine Learning classification — automatic categorisation of incoming documents.

·       Intelligent extraction — structured data (partner, dates, numbers, amounts, VAT, ICE, lines).

·       Human validation — confidence scoring per field and a dedicated validation centre.

·       Documentary workflow & business automation — rule-based creation of Dolibarr objects.

·       Intelligent GED (DMS) — classification, archiving, versioning and access rights.

·       Intelligent search — full-text, semantic and image-content search.

·       Audit & GDPR traceability — complete action journal per document.

The architecture is open and compatible with external providers: OpenAI API, Azure AI Vision, Google Document AI, AWS Textract and Tesseract OCR.

3. Key features

·       Multi-source import: manual upload, scanner, incoming email, API, DMS, watched folder, mobile and cloud storage.

·       Supported formats: PDF, JPG, PNG, TIFF, Word and Excel.

·       Custom OCR models with extraction zones, fields and validation rules.

·       Confidence score per document and per field (green >= 95% auto, orange 80-95% review, red < 80% correction).

·       Validation centre: side-by-side OCR text and extracted data with inline correction.

·       AI assistant: document summary, anomaly detection, contract obligations, document comparison and Q&A.

·       Modern dashboard and decisional reporting with CSV export and printing.

·       First-activation demonstration dataset with one-click install / removal.

4. Business objects

The module ships 15 fully integrated business objects:

5. Documentary workflow

A typical supplier-invoice workflow illustrates the end-to-end pipeline:

·       Reception of the supplier invoice (upload, email, scanner...).

·       OCR recognition of the document pages.

·       Intelligent extraction of the fields (supplier, number, date, HT, VAT, TTC, ICE).

·       AI control and classification of the document.

·       Human validation in the validation centre (only for orange/red confidence).

·       Automatic creation of the Dolibarr supplier invoice.

·       Notification of the accounting team and archiving of the document.

6. Artificial intelligence

The AI layer provides document understanding, automatic classification, intelligent extraction and an assistant able to summarise documents, detect anomalies and inconsistencies, extract contractual obligations, compare documents, generate reports and answer documentary questions.

Each extraction carries a global and per-field confidence score, driving the automatic-versus-manual validation decision and continuous quality measurement.

7. API & extensibility

The provider layer abstracts the OCR/AI engine so the module can operate with a local engine (Tesseract) or delegate to cloud providers through their APIs:

·       OpenAI API (GPT Vision) for document understanding and Q&A.

·       Azure AI Vision for OCR and layout analysis.

·       Google Document AI for structured extraction.

·       AWS Textract for tables and forms.

·       Tesseract OCR for on-premise, privacy-preserving recognition.

Each provider is configured with its endpoint, API key, model and supported languages, and can be enabled or set as default independently. The module also integrates with the Dolibarr REST API and webhooks.

8. Dolibarr integrations

Documents processed by the module can automatically feed the native Dolibarr objects and modules:

·       Business documents: quotes, customer/supplier orders, customer/supplier invoices, shipments, contracts.

·       Modules: Third parties, Products, Stocks, Warehouses, Accounting, Projects, HR, GED/DMS, CRM, Agenda.

·       Technical: REST API and webhooks for inbound and outbound integration.

9. Security & compliance

·       Granular permission groups (documents, models, AI, workflows, configuration, search, audit, reporting).

·       Complete audit trail and modification history per document.

·       Full traceability for GDPR compliance.

·       Document access control and user-based rights.

© 2026 DoliResources — www.doliresources.com