System Configuration & AI Model Settings

YOLOv8 Segmentation Thresholds • Monocular Depth Matrix Scaling • Groq LLM Runtime

COMPUTER VISION & DEPTH MATRIX PIPELINE

YOLOv8-Seg Instance Confidence Threshold:85%

50% (High Recall)85% (Optimal)99% (Strict Precision)

Depth Anything Monocular Scale Factor:1.20x

0.80x (Shallow)1.20x (Standard)2.00x (Deep Extrusion)

NIRMAN COPILOT & LLM RUNTIME

Active Groq LLM Inference Model:

Indian Road Congress (IRC) Governance Standard:

SWYTCHCODE DURABLE WORKFLOW & AGENTIC API ENGINE

Swytchcode v2.4 Active

Durable Retry Policy

Exponential Backoff (3 Retries)

State Persistence

Enabled (Session Key Sync)

Swytchcode CLI OpenAPI Spec

CODEMATE AI SDLC QUALITY & OWASP SECURITY AUDITOR

OWASP Security Rating

A+ (0 Critical Vulns)

TypeScript Strictness

100% Type Coverage

Automated PR Review Agent

Passing (GitHub Actions)

CodeMate Audit API