Step 3: Generate your README

Run the full generator and review what each section produces.

Generate all sections

Sample projectLocal directoryGitHub repo

python readme_generator.py --sample --output README_sample.md

python readme_generator.py --input /path/to/your/project --output README.md

python readme_generator.py --input owner/repo --output README.md

The script generates each section in order, prints progress, and writes the assembled README to the output file.

Expected output:

📄 README Generator
==================================================

🔍 Analyzing ./my-project...
✓ Project: my-project
✓ Language: Python
✓ License: MIT
✓ Dependencies: click, requests (+2 more)

✍️  Generating description section...
✓ Description section complete

✍️  Generating installation section...
✓ Installation section complete

✍️  Generating usage section...
✓ Usage section complete

✍️  Generating contributing section...
✓ Contributing section complete

📝 Assembling README...
✓ README written to README.md

✅ Done! Review the output before publishing.

Generate specific sections only

Use --sections to regenerate just one or two sections instead of all four:

# Regenerate only the usage section
python readme_generator.py --input ./my-project --sections usage --output README.md

# Regenerate description and installation
python readme_generator.py --input ./my-project --sections description,installation --output README.md

This is useful when you want to keep sections you've already edited and only refresh specific ones.

Review each section

Open the generated README and evaluate each section critically.

Description section

AI is generally good at this. Check:

• Does it accurately describe what the project does?
• Is the feature list realistic (not invented)?
• Is the tone appropriate for your audience?

Installation section

AI infers install commands from dependency files. Verify:

• Are the prerequisite versions correct?
• Does pip install -r requirements.txt (or equivalent) actually work?
• Is the quickstart example syntactically correct?

Always test installation steps

AI generates plausible-looking commands based on patterns, but it cannot run them. Every installation step must be tested manually before publishing.

Usage section

AI generates examples based on your entry point files. Verify:

• Do the example commands actually work?
• Are function signatures and argument names accurate?
• Is expected output shown correctly?

Contributing section

This section is largely boilerplate and is usually safe as-is. Check:

• Is the license type correct?
• Are the contribution steps appropriate for your project?

What to keep, what to rewrite

AI generates well	Write yourself
Project descriptions	Specific version requirements
Feature bullet lists	Accurate code output/screenshots
Boilerplate contributing text	Security or authentication notes
License notices	Roadmap or known limitations
General install patterns	Environment-specific instructions

---

Next step: Refine and customize →