CAD Intelligence Extraction Agent using LangChain and Agno Framework
3D CAD models alone lack integrated manufacturing metadata, forcing manual interpretation of STEP files and bill-of-materials. Build an engineering intelligence pipeline using LangChain and Agno to parse CAD assemblies, extract design constraints, and automatically output structured manufacturing specs with 95% accuracy.
What you are building
The core problem, expected build, and operating context for this challenge.
Construct an agentic engineering pipeline with LangChain and Agno to analyze raw STEP files and generate production-ready bill of materials and machining instructions.
How work is evaluated
Measures tolerance extraction correctness and BOM material accuracy from sample CAD files.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Material Detection Accuracy
Checks if the extracted material matches target alloy requirements.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
spec_completeness
Percentage of required manufacturing attributes extracted. • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Parse 3D CAD geometry metadata from STEP files using python-occ and Agno agents
Build a LangChain multi-prompt workflow to transform geometric dimensions into CNC machining steps
Validate extracted component tolerances against manufacturing capability parameters
Reference links and supporting material
Extracted STEP file metadata headers and geometrical properties from open industrial fixture assemblies.
Challenge setup
Measures tolerance extraction correctness and BOM material accuracy from sample CAD files.
Challenge input
STEP file metadata string or path
LangChain
Handles structured chaining and dynamic template execution for metadata conversion.
Agno
Provides stateful memory and light agent structures for CAD hierarchy tracking.
Sourcegraph Amp
Accelerates codebase refactoring and AST synthesis for engineering schema updates.
Evaluated output
JSON containing part_name, material, iso_tolerance, and machining_operations
- Checks if the extracted material matches target alloy requirements.
- Percentage of required manufacturing attributes extracted. • target: 0.95 • range: 0-1
- Benchmark: Engineering Design Extraction Evaluation 2026
- Spec Completeness target: 0.95
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Langchain · Required
- LangChain · Optional
- Agno · Optional
- Langchain · Required
- LangChain · Optional
- Agno · Optional
Evaluation contract
- Material Detection Accuracy · Weight 1
- spec_completeness · Weight 1
Metrics
- spec_completeness · Target 0.95
Benchmarks
- Engineering Design Extraction Evaluation 2026
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: STEP file metadata string or path
Output: JSON containing part_name, material, iso_tolerance, and machining_operations
Execution harness: Python
Evaluation tools
- LangChain · LLM Orchestration Framework
- Agno · Agent System Architecture
- Sourcegraph Amp · AI Engineering Tooling
Recipe freeze marker
No recipe freeze time is recorded.