README.md

CodePapi Backend

Overview

NestJS API server that orchestrates local LLM inference via Ollama and LangChain. Provides endpoints for code translation, review, and bug detection.

Note: This is part of an experimental hobby project. Code quality and performance optimizations are secondary to learning and experimentation.

Project Setup

npm install

Running the Application

# Development mode
npm run start:dev

# Production mode
npm run start

The API runs on http://localhost:3000

Endpoints

• POST /converter/translate — Translate code between languages
• POST /converter/review — Get AI feedback on code
• POST /converter/fix — Get AI bug fixes with diff

All requests expect JSON bodies with code and language parameters.

Architecture

• Controller: converter.controller.ts — HTTP request handling
• Service: converter.service.ts — AI prompt orchestration via LangChain
• Module: converter.module.ts — Dependency injection setup

Key Dependencies

• @langchain/ollama — LLM integration with Ollama
• @nestjs/core — Framework
• @nestjs/common — Common utilities

Performance Notes

• AI responses take 10-90+ seconds depending on code size and hardware
• Requests are blocking (no queue system)
• Suitable for local development, not production use
• Single model instance (Qwen2.5-Coder)