neurotidy 0.1.0

AI-Powered Python & Deep Learning Code Analyzer — Explain, Analyze, Optimize, and Debug your code

NeuroTidy

AI-Powered Python & Deep Learning Code Analyzer

NeuroTidy analyzes your Python and PyTorch/TensorFlow code using AI. It can explain, analyze, optimize, and debug your code — all from the terminal.

Install

pip install neurotidy

Usage

# Explain code for beginners
neurotidy explain train.py --mode beginner

# Static code analysis
neurotidy analyze model.py

# DL-specific optimization suggestions
neurotidy optimize train.py

# Debug an error message
neurotidy debug --error "RuntimeError: mat1 and mat2 shapes cannot be multiplied"

# Inline code
neurotidy explain --code "def add(a, b): return a + b" --mode beginner

Features

• Explain — AI-powered code explanations at beginner, intermediate, or advanced level
• Analyze — Static analysis with Python best-practice rules (PY001–PY010)
• Optimize — Deep learning specific optimizations (NT001–NT027) with RAG-enhanced analysis
• Debug — Paste an error message, get root cause analysis and step-by-step fixes

How It Works

NeuroTidy calls a serverless AWS backend:

• API Gateway → Lambda → Amazon Bedrock (Claude AI)
• Results cached in DynamoDB, stored in S3
• Zero infrastructure to manage — just pip install and go

Configuration

By default, NeuroTidy connects to the hosted API. To use a custom endpoint:

export NEUROTIDY_API_ENDPOINT=https://your-api.execute-api.us-east-1.amazonaws.com/prod

Built by

ERROR-108 — AWS Hackathon 2026

License

MIT