DevicesSensorsSecurityMobile app
Completed projectEU-fundedSmart Home / IoT

OP “Innovation and Competitiveness” 2014–2020

Implementation of an Innovative Smart Home Platform

Since 7 July 2017, CS Development Ltd. has been implementing grant contract No. BG16RFOP002-1.001-0208-C01 for the project “Implementation of an Innovative Smart Home Platform”.

The main objective of the project is to provide investment support for the technological capacity required to bring an innovative product to market — the commercial launch of the Smart Home platform.

The total project value is EUR 196,947.59, of which EUR 150,664.90 is European and EUR 26,587.92 is national co-financing. The company’s own co-financing amounts to EUR 19,694.76.

The project end date is 7 October 2018.

Project details
ProgrammeOP “Innovation and Competitiveness” 2014–2020
Grant contract No.BG16RFOP002-1.001-0208-C01
BeneficiaryCS Development Ltd. (СИ ЕС ДИВЕЛЪПМЪНТ ООД)
Period7 July 2017 – 7 October 2018
Total valueEUR 196,947.59
European co-financingEUR 150,664.90
National co-financingEUR 26,587.92
Own co-financingEUR 19,694.76

Amounts converted to euro at the fixed rate EUR 1 = BGN 1.95583. Original amounts: BGN 385,196.00 in total; BGN 294,674.94 European, BGN 52,001.46 national and BGN 38,519.60 own co-financing.

Official text (in Bulgarian)

ОП „Иновации и конкурентоспособност“ 2014-2020

От 07 юли на 2017 г. СИ ЕС ДИВЕЛЪПМЪНТ ООД изпълнява договор за БФП № BG16RFOP002-1.001-0208-C01 с наименование “Внедряване на иновативна платформа Smart Home”.

Основната цел на проекта е осигуряване на инвестиционна подкрепа за осигуряване на технологична обезпеченост за пазарното внедряване на иновативен продукт - пазарна реализация на платформа Smart Home.

Общата стойност на проекта е 196 947,59 евро, от които 150 664,90 евро европейско и 26 587,92 евро национално съфинансиране. Собственото съфинансиране е в размер на 19 694,76 евро.

Крайната дата на проекта е 07 октомври 2018 г.

DocumentsAI agentsData
AI platformInvoices & documentsLogistics

PdfAgent — AI Platform for Invoice and Document Processing

PdfAgent is an intelligent document processing system that automates the extraction, validation and routing of supplier invoices and logistics documents. Built for the transport and logistics industry, it turns unstructured PDFs — invoices, CMR waybills and e-archive documents — into clean, structured, database-ready data with minimal human intervention.

Key capabilities

  • End-to-end automation — watches an input folder, classifies each document, extracts the key invoice fields (vendor, tax IDs, dates, amounts, currency, IBAN, tour references), validates them and routes the file to Success, Error or NonInvoice, with a full audit trail.
  • Multi-source recognition — combines embedded PDF text, OCR and a Vision-Language Model, then cross-checks the result against deterministic and country-specific rules.
  • Grounded, not guessed — every value is validated against contractor and transport data (via MCP services), QR-code payloads (Turkish e-Arşiv, Polish KSeF) and a recognition knowledge base. Hallucinated data is rejected rather than passed on.
  • Configurable pipeline — a JSON-defined processing pipeline, hierarchical prompt management and per-step artifacts (prompts, raw model output, step JSON) for full traceability.
  • Enterprise integrations — PostgreSQL, optional MinIO object storage, links to ERP and contractor/transport systems, and QR-code verification.

Local AI and data privacy

All AI recognition runs on a locally hosted model inside the company’s own infrastructure. No document, image or extracted value ever leaves the customer’s environment, which makes the system suitable for financial documents and sensitive supplier data where cloud AI services are not an option.

Business value

  • Manual entry becomes review-only — accountants check only the documents the system flags as unclear or suspicious.
  • Fewer errors and overpayments — every invoice is matched automatically to the right partner and transport task, and amounts, currencies and IBANs are cross-checked before they reach accounting.
  • Full accountability — each document keeps an evidence trail (source text, model output, validation decision), so any figure can be traced to its source in seconds.
  • Works day and night — invoices arriving by email are processed automatically, so records are ready in the ERP system when the accounting team starts the day.
  • Scales with the business — the same pipeline handles dozens or hundreds of documents a day without additional headcount.
New project?

Which tasks should we automate for you?

Tell us how you work today — we will show you where AI agents and automation save the most time.