KabyTech — Freight Document Intelligence for Thai Logistics
Backend Featured

KabyTech — Freight Document Intelligence for Thai Logistics

A B2B SaaS platform that turns multi-modal freight paperwork — AWB, B/L, Sea Waybill, Rail and Road consignments, Delivery Orders, and Freight Manifests — into validated, structured JSON with per-field confidence scoring, sub-2-second latency, and native Thai-language support.

Overview

KabyTech is a production document-intelligence platform (kabytech.com) for Thai freight forwarders, customs brokers, ground handlers, and cargo airlines. It parses seven freight document types end-to-end, validates them against IATA, customs, and Thai regulatory rules, and ships an Operations Portal and a full Freight ERP on top of the same intelligence layer.

Key Features

  • Seven document types end-to-end: Air Waybill (AWB), Bill of Lading (B/L), Sea Waybill (SWB), Rail Consignment (RCN), Road Consignment (CMR), Delivery Order (DO), and Freight Manifest (MFT)
  • Deep field extraction — 120+ fields per AWB, 45+ per B/L, 35–50 per document — with semantic Cargo-IMP understanding rather than regex
  • Native Thai-language reading of Thai script, numerals, and mixed Thai/English documents — no external translation in the pipeline
  • Live validation against IATA codes, HS codes, Thai airport codes (VTBS, VTBD, VTSP), and Thai tax IDs cross-checked with customs reference data
  • Per-field confidence scoring (98.7%–99.5% on production traffic) and structured error reports for downstream review
  • Three product tiers: a Document Intelligence REST API, an Operations Portal with team workflow / RBAC / batch processing, and a full Freight ERP with auto-reconciliation
  • Native integrations with CargoWise One, SAP Transportation Management, Oracle WMS, Thai NSW e-Customs EDI, and IATA CASS billing
  • SDKs for Python, Node.js, PHP, and Java with webhook ingest for batch EDI feeds

Architecture

Component Details
Document Intelligence Engine Semantic NLP/NLU parser that understands the 29-segment IATA Cargo-IMP FWB/16 structure (SHP, CNE, AGT, FLT, RTG, RTD, CVD, OCI, …) and the equivalent layouts for B/L, SWB, RCN, CMR, DO, and MFT
Validation Layer Enforces IATA conditional rules, HS-code validity, Thai airport / airline / tax-ID cross-references, and CIM/SMGS + GMS/CBTA rules for cross-border rail and road
Confidence & Audit Every extracted field carries a confidence percentage; full reconciliation history is persisted for audit, replay, and partner onboarding
REST API HTTP endpoints that accept raw documents and return structured JSON with shipment, routing, weights, parties, and linked house waybills
Operations Portal Next.js web console for upload, review, batch processing, team workflow, RBAC, and export to Excel/CSV/CargoWise/SAP/e-Customs
Freight ERP Integrated TMS layered on the intelligence engine — auto-reconciliation, IATA CASS connectivity, multi-branch and multi-currency support
Integrations Native connectors for CargoWise One, SAP TM, Oracle WMS, Thai NSW e-Customs EDI, IATA CASS, and direct feeds from Thai Airways Cargo and Bangkok Airways
Production Operated at kabytech.com with a 99.9% uptime SLA, p99 latency under 2 seconds, SOC 2 Type II certification, ISO 27001 in progress, and PDPA compliance

Key Metrics

Document types

7 (AWB, B/L, SWB, RCN, CMR, DO, MFT)

Field extraction accuracy

97%+ overall · 98.7–99.5% per field

API response (p99)

<2s per document

Uptime SLA

99.9%

Customer reference throughput

800+ MAWB/HAWB per day

Tech Stack

Next.jsTypeScriptNode.jsNLP/NLUREST APIIATA Cargo-IMPPostgreSQLDocker

Details

Category Backend
Technologies 8
Status Published
Featured Yes

Gallery

Related Portfolio