Foreign trade / customs In progress

Finding TARIC goods codes via RAG

The TARIC database as a local RAG: matching import/export goods codes are found from a plain description of the goods — replacing tedious manual search through the nomenclature.

The challenge

In foreign trade, every good needs the right TARIC goods code — an extensive, nested nomenclature whose manual lookup takes experience and a lot of time.

How the stack solves it

LightRAG indexes the TARIC database locally and matches a plain-language goods description to the right codes, with Open WebUI providing the chat interface. The customs clerk asks the question and picks the correct code from the suggestions — the nomenclature never leaves the building.

The workflow in the rzfz.ai stack

The TARIC classification workflow in the rzfz.ai stack (Dify) — from goods input to the customs code
The RAG pipeline behind TARIC classification: retrieving the matching customs-tariff passages before the local model decides the code
The TARIC assistant in chat: a product description goes in, the answer names the customs code together with its reasoning

Click to enlarge.

How the components work together

Customs clerk extern
Open WebUI · chat
Local LLM
LightRAG · TARIC index
Matching sections Suggestions in chat

The result

The right goods codes are found in seconds instead of after a lengthy search — the classification work stays entirely local.

Something like this for your company?

Describe your case in the use-case finder, or talk to us about an implementation in the configurator.

Open the configurator →

Open the use-case finder → See consulting →