A technical engineering reference for building, maintaining, and scaling automated customs-compliance pipelines.

🌐 www.customs-brokerage.org

Customs clearance is an engineering problem long before it is a paperwork problem. Every duty figure a broker files traces back through a chain of deterministic, auditable transformations: a commercial description resolved to a Harmonized System code, a temporally-versioned tariff schedule, an origin determination, a duty formula, and a wire-format transmission to CBP. When any link in that chain is non-deterministic, duty leaks, filings get rejected, and a Focused Assessment finds gaps no one can reconstruct.

This site is a deep, code-first reference for the Python ETL teams, customs brokers, and trade-compliance officers who build those pipelines. Every article leads with runnable Python 3.10+ (exact Decimal money, structured logging, asyncpg/PostgreSQL, idempotent upserts) and anchors each pattern to the specific regulatory obligation it satisfies — WCO HS 2022 nomenclature, HTSUS (USITC), CBP ACE decimal-precision and rounding rules, ISF 19 CFR 149, and Post-Summary Correction workflows.

The reference is organised into three connected areas, each with in-depth guides and focused implementation walkthroughs:

Treating the Harmonized System as a versioned, temporally-aware data layer.

• Temporal HTS schedule database design and point-in-time resolution with recursive CTEs
• Idempotent tariff-update ingestion pipelines and Polars memory profiling for large HTSUS releases
• Rule-of-origin engines: regional value content over BOM hierarchies, USMCA tariff-shift testing
• Duty-formula frameworks: ad valorem, specific, and compound rates with CBP half-up rounding
• Fallback routing for unmapped codes, plus RBAC, envelope encryption, and audit-grade data isolation

Turning heterogeneous commercial documents into deterministic, schema-bound payloads.

• Commercial-invoice PDF extraction (pdfplumber vs camelot vs tabula), multi-currency handling with Babel
• Packing-list normalization and net-vs-gross weight reconciliation
• OCR drift correction and confidence-threshold calibration for HS digit fields
• Multi-language and CJK invoice parsing with Unicode NFC normalization
• Exponential backoff, dead-letter queues, and high-volume async batch processing

Assembling and transmitting the entry to CBP, and keeping it defensible through liquidation.

• Entry summary (CBP 7501) line assembly and ACE decimal-precision validation
• Importer Security Filing (ISF 10+2) data assembly and late-filing / liquidated-damages handling
• ABI/ACE rejection handling: decoding X12 824 application advice and idempotent resubmission
• Post-entry amendment workflows: Post-Summary Corrections and duty-refund reconciliation

• Python ETL / data engineers building customs and trade-compliance data pipelines
• Customs brokers and licensed professionals automating classification and filing
• Trade-compliance officers who need audit-ready, reconstructable workflows
• Logistics developers integrating tariff, valuation, and ACE transmission systems

A static site generated with Eleventy, authored in Markdown with hand-crafted inline SVG diagrams, KaTeX for duty-formula notation, structured data (JSON-LD), and a full accessibility and performance budget. It is deployed on Cloudflare Workers.

npm install
npm run build      # build the static site into _site/
npm run serve      # local preview on :8080
npm run deploy     # build + deploy to Cloudflare

Content and code samples are published as an educational engineering reference. The code patterns are illustrative — validate against the current WCO, USITC, and CBP instruments before relying on them for live customs filings.

Built for the people who keep goods moving across borders — explore the full reference →