eDocify platform overview
eDocify is a Bitlogika document operations platform for companies that need one controlled way to receive, recognize, verify, approve, export, and archive business documents. It is designed as a product family rather than one generic OCR screen: accountants, verifier teams, records managers, auditors, and integration owners each get their own workspace.
Product family
| Product | Primary users | Main outcome |
|---|---|---|
| Accounting Workspace | Accountants, AP teams, accounting service firms | Fast invoice, receipt, approval, and ERP export workflow. |
| Enterprise IDP / OCR Operations | Document factories, verifier teams, BPO operations | High-volume OCR, AI verification, quality governance, SLA control, and multi-document operations. |
| E-document Archive | Records managers, auditors, compliance teams | Searchable, retention-aware, audit-ready long-term document archive. |
Demo and documentation evidence
The screenshots in this documentation are not memory mockups. They are captured from the local eDocify app with English UI, product-specific demo sandbox sessions, and role-specific navigation. The generated manifest is stored at /img/edocify/role-workspaces/manifest.json and records the product, role, route, visible menu items, and detected UI errors for every screenshot.
Platform flow
flowchart LR
A["Intake sources"] --> B["Preparation: split, merge, rotate, security checks"]
B --> C["OCR / AI provider routing"]
C --> D["Field processing and validation"]
D --> E["Verification workbench"]
E --> F["Approval workflow"]
F --> G["ERP export or archive"]
E --> H["AI Learning and Quality Engine"]
H --> C
What makes the platform enterprise-ready
Role-specific workspaces. A verifier does not see the same cockpit as a system administrator. An archive viewer does not see AI Learning. A sandbox admin stays inside the selected product sandbox.
Multi-engine OCR strategy. eDocify supports Azure Document Intelligence, Mistral OCR, OpenAI-style AI providers, local Tesseract, local PaddleOCR, deterministic invoice rules, and hybrid BYOK routes.
Quality governance. OCR quality is treated as a release process: golden datasets, field-level accuracy, provider bake-offs, confidence calibration, and release gates.
Audit-first operations. Intake, verification, approval, export, archive, sandbox, and admin actions are designed to create traceable audit evidence.
Azure-ready, VPS-friendly. The architecture is prepared for Azure hosting, but can also run on a controlled VPS for early public demos and pilots.