churn-cli 0.1.1

Read your codebase. Generate interview questions.

Churn

A personal developer CLI tool that reads your codebase and generates interview questions based on your actual project.

---

What It Does

Churn scans a local project directory, identifies the most frequently edited files using git log, compresses the code, and sends it to the Gemini API. It returns 15 interview questions with detailed answers across 3 experience levels, saved as a markdown file.

---

Current State

This is a personal project under active development. It is not published to PyPI yet.

What works today:

• Local project directory scanning (recursive, all subfolders)
• Git-based important file detection via git log --stat
• Code compression — strips comments and blank lines
• Gemini API integration (google-genai)
• Structured JSON output parsed into a markdown file
• 3 experience levels: Fresher / Mid / Senior — 5 questions each

---

Usage

Requirements: Python 3.8+, a Gemini API key from aistudio.google.com

# Clone the repo
git clone https://github.com/dhawalshankar/churn
cd churn

# Create and activate virtual environment
python -m venv churn-env
churn-env\Scripts\activate  # Windows
source churn-env/bin/activate  # Mac/Linux

# Install dependencies
pip install google-genai click python-dotenv

# Add your Gemini API key
echo GEMINI_API_KEY=your_key_here > .env

# Run
python main.py ./your-project-path

Output is saved to churn-output.md in the current directory.

---

Output Structure

## FRESHER
Q1: What does this project do?
A: ...

## MID
Q1: Why did you choose this database over alternatives?
A: ...

## SENIOR
Q1: What are the security risks in your current architecture?
A: ...

---

Files Scanned

Included: .py .js .ts .tsx .jsx .java .go .cpp

Ignored: node_modules/ .git/ dist/ build/ __pycache__/ .next/ venv/ churn-env/

---

Privacy

• No code is stored at any point
• Everything runs locally — no backend server
• Code is processed in memory, sent to Gemini, and wiped immediately
• Gemini API does not use API request data for model training

---

Tech Stack

• Python 3.12
• google-genai — Gemini API SDK
• click — CLI
• python-dotenv — environment variables

---

Roadmap

• PyPI publish — pip install churn
• GitHub public repo URL support
• VS Code Extension
• Custom focus flag — --focus security
• PDF export

---

License

This is a personal project. All rights reserved.

© 2026 dhawalshankar