doced 3.0.1

Automate software documentation using AI (Gemini & K2 Think).

doced

Automate software documentation using AI reasoning.

doced is a CLI tool that scans your codebase, interviews you with intelligent questions, and generates comprehensive documentation including README, tutorials, architecture diagrams, and PDFs.

Features

• 🔍 Smart Codebase Scanning - Respects .gitignore and intelligently filters files
• 🤖 Multi-AI Approach - Uses Gemini for architecture mapping and K2 Think for deep reasoning
• 💬 Interactive Interview - AI generates targeted questions based on your code
• 📚 Comprehensive Docs - Generates README, tutorials, architecture docs, and PDFs
• 🎨 Customizable Styles - Choose between Corporate/Creative tones and Brief/Detailed levels
• 📊 Mermaid Diagrams - Automatically generates architecture and flow diagrams

Installation

Recommended (Auto-handles PATH)

pip install pipx
pipx install doced
pipx ensurepath

Alternative

pip install doced

Development Setup

# Clone the repo
git clone https://github.com/qmamab/doced.git
cd doced

# Install dependencies
pip install -r requirements.txt

Usage

1. Initialize (First Time Setup)

doced init

Sometimes that doesnt work on mac's, so try

python3 -m doced init

You'll be prompted for:

• Gemini API Key
• K2 Think API Key
• Documentation Style (Corporate/Creative)
• Detail Level (Brief/Detailed)

2. Generate Documentation

# In your project directory
doced create

The tool will:

1. Scan your codebase
2. Analyze architecture with Gemini
3. Generate intelligent questions with K2 Think
4. Interview you interactively
5. Create comprehensive documentation in docs/
   • README.md
   • TUTORIAL.md
   • ARCHITECTURE.md
   • Documentation.pdf

API Configuration

Gemini API

• Model: gemini-2.0-flash-latest
• Get your key: https://makersuite.google.com/app/apikey

K2 Think API

• Model: MBZUAI-IFM/K2-Think
• Endpoint: https://api.k2think.ai/v2/chat/completions
• Rate Limit: 20 RPM
• Get your key: Contact K2 Think team

Project Structure

doced/
├── pyproject.toml        # Dependencies and metadata
└── src/
    └── doced/
        ├── main.py       # CLI and orchestration
        └── scanner.py    # Code scanning logic

Configuration

Your config is stored at ~/.doced/config.json. You can re-run doced init to update it.

Troubleshooting

Import Errors

Make sure you're in the virtual environment:

source .venv/bin/activate

API Errors

• Verify your API keys are correct
• Check rate limits (K2 Think: 20 RPM)
• Ensure you're using the correct model names

License

MIT