HS-Code Search Logic &
Intelligent Pipeline
TENSIG utilizes a multi-layered extraction architecture to decode complex import descriptions into precise Harmonized System codes.
account_tree Execution Flow
Input Processing
Sanitizing raw user strings and metadata
Direct Code Detection
Regex-based scanning for 4-10 digit sequences
Semantic Search
PrimaryVector-based intent & context matching
Full-Text Search
BM25 matching against official TARIC labels
Fuzzy Normalization
Correcting typos and variant spelling
Language Fallback
Cross-referencing EN descriptors if NL fails
The Search Methodology
Direct Match
Instantly validates exact codes. Reduces processing time by 85% for known shipments.
Vector Embedding
The Semantic engine understands "circular cooling device" is likely a Fan (8414.59).
Normalization
Converts plural/singular and handles regional variations (e.g., 'Elektrisch' vs 'Electrisch').
Full-Text Index
Scans legal notes and exclusionary clauses to ensure compliance precision.
Language Bridging: NL ↔ EN
Official TARIC documentation is primarily maintained in English. TENSIG bridges the gap by translating Dutch technical terms into high-fidelity English synonyms before querying the semantic vector space.
Pro-Tip: Semantic search works best with descriptive nouns rather than technical verbs.
Automated Compliance
Every search is cross-referenced with the latest EU import regulations to minimize clearing delays.