Automated Crude Loading Manifest Validation Pipeline with Pydantic AI
Before an oil tanker can depart terminal waters, its bill of lading, customs clearance, and quality manifest must match export quota allowances down to the net standard volume (NSV). Build an automated validation agent with Pydantic AI. The agent receives manifest inputs, checks them against interim 3-month export license limits, and validates volume units across metric tons, long tons, and standard US barrels.
What you are building
Develop an export loading manifest validation system using Pydantic AI schema enforcement.
How work is evaluated
Evaluates loading manifest volumetric compliance against quota.
Datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
quota_exceeded_false
Ensure manifest within quota gets valid clearance.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
volumetric_calculation_precision
Precision of calculated barrels from metric tons. • target: 0.99 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Learning goals
Use Pydantic AI to enforce multi-field numerical consistency.
Implement automatic volumetric conversions for crude oil quantities.
Generate export clearance tokens for valid manifests.
Resources
50 manifest records containing tonnage, API gravity, and quota limits.
Challenge setup
Evaluates loading manifest volumetric compliance against quota.
Challenge input
JSON with metric_tons, api_gravity, quota_max_bbls
Pydantic AI
Type-safe validation engine for structured agent outputs.
Evaluated output
JSON with calculated_bbls, quota_exceeded, clearance_valid
- Ensure manifest within quota gets valid clearance.
- Precision of calculated barrels from metric tons. • target: 0.99 • range: 0-1
- Benchmark: Pydantic-Validation-Bench
- Volumetric Calculation Precision target: 0.99
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
- Pydantic AI · Required
- VAPI · Optional
- Langfuse · Optional
- Langfuse · Optional
- Pydantic AI · Required
Evaluation contract
- quota_exceeded_false · Weight 1
- volumetric_calculation_precision · Weight 1
Metrics
- volumetric_calculation_precision · Target 0.99
Benchmarks
- Pydantic-Validation-Bench
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON with metric_tons, api_gravity, quota_max_bbls
Output: JSON with calculated_bbls, quota_exceeded, clearance_valid
Execution harness: Python
Evaluation tools
- Pydantic AI · Schema Engine
Recipe freeze marker
No recipe freeze time is recorded.