mcop

mcop is the Python distribution for the Meta-Cognitive Optimization Protocol, a reasoning framework for multi-modal hypothesis generation, recursive refinement, and evidence-grounded synthesis.

The package exposes:

• A general-purpose reasoning engine.
• Domain adapters for general, medical, and scientific workflows.
• A command-line interface for interactive use and scripted runs.
• Structured outputs with confidence, grounding, evidence, and alternatives.

Install

pip install mcop

Optional extras:

pip install mcop[llm]
pip install mcop[dev]

Quick Start

Solve a problem directly

from mcop import solve

solution = solve("What causes climate change?")
print(solution.content)
print(f"Confidence: {solution.confidence * 100:.1f}%")
print(f"Grounding index: {solution.grounding_index:.2f}")

Work with the engine explicitly

from mcop import MCOPEngine, Problem

engine = MCOPEngine()
problem = Problem(description="Your problem here")
solution = engine.solve(problem)
print(solution.content)

Use a domain adapter

from mcop.domains import MedicalDomainAdapter, PatientPresentation

adapter = MedicalDomainAdapter()
presentation = PatientPresentation(
    chief_complaint="Chest pain",
    symptoms=["chest pain", "shortness of breath"],
)
problem = adapter.create_patient_problem(presentation)
solution = adapter.solve(problem)
print(adapter.format_differential_diagnosis(solution))

Command Line Interface

mcop solve "What are the causes of inflation?"
mcop solve --domain medical "Patient with fever and cough"
mcop interactive
mcop info

What the Package Returns

Each solution includes the primary response plus supporting metadata such as:

• Confidence score.
• Grounding index.
• Evidence chain.
• Alternative solutions.
• Key uncertainties.

Project Resources

• Repository: Kuonirad/MCOP-Framework-2.0
• Usage guide: mcop_package/USAGE_GUIDE.md
• API reference: mcop_package/API_REFERENCE.md
• Project structure: mcop_package/PROJECT_STRUCTURE.md
• Changelog: CHANGELOG.md

Notes

• The Python package has no required runtime dependencies.
• Medical and scientific adapters are decision-support examples and do not replace professional judgment.
• Trusted publishing setup for PyPI is documented in TRUSTED_PUBLISHING_SETUP.md.

License

Apache License 2.0 (Apache-2.0) — see LICENSE and NOTICE.md for terms. Versions originally released under MIT remain available under MIT (see LICENSE-MIT-LEGACY).