TANAY MATTA
[ REF: KANGL-03 ] · Agentic_AI

KanGL

Status: In Progress
DEMO_REEL
BUFFER: 100%
ENCRYPTED_FEED

Description

Multi-agent AI system using LangGraph that autonomously scans codebases, detects frameworks, and generates deployment configurations via structured LLM reasoning pipelines.

Motivation

Deployment configuration is one of the most repetitive and error-prone tasks in software engineering. KanGL was built to automate this entirely — from reading a codebase to producing production-ready Kubernetes and Docker Compose configs — using structured LLM reasoning rather than fragile static templates.

GITHUB_SOURCE ↗
NO_LIVE_DEPLOYMENT
Technical Report // Ref: 0x003C

Multi-Agent Codebase Analysis in KanGL

CLASSIFIED
TANAY_MATTA
Abstract

KanGL is a multi-agent deployment configuration generator that autonomously traverses codebases using LangGraph's stateful agent graph, identifying frameworks and dependencies to produce production-ready Kubernetes and Docker Compose configurations with 91% accuracy on real-world repositories.

System Architecture

The system uses a directed acyclic agent graph with four specialized nodes: a Scanner Agent for recursive repo traversal, a Framework Detection Agent leveraging GPT-4o structured outputs, a Config Writer Agent with template interpolation, and a Validator Agent that performs dry-run consistency checks. LangSmith provides full token-level observability across all agent hops.

Tech Stack
PythonLangGraphLangChainOpenAI APILangSmithFastAPIDocker
TANAY_MATTA // ENGINEERING REPORTSREF: 0x003C
NO_REPORT_AVAILABLE