Skip to content

Tutorial overview

Build your first release notes automation using AI coding agents. This tutorial shows you how to automate release notes generation from GitHub commits.

Build these components

  • Working release notes automation script
  • Documented manual process
  • Customized AI categorization prompts
  • Iteration and improvement workflow
  • Reusable pattern for future automations

Follow these steps

1. Setup environment

Time: 15 minutes

Install dependencies, create configuration files, and verify everything works.

  • Install Python and dependencies
  • Set up API keys
  • Run your first test

2. Document your process

Time: 30 minutes

Write down your current manual release notes workflow. This is the most important step.

  • Complete the workflow worksheet
  • Define your categorization standards
  • Identify exclusion rules
  • Document output format

Your documented process becomes the foundation for your automation prompt.

3. Configure APIs

Time: 10 minutes

Set up GitHub and AI provider API access.

  • Create GitHub token
  • Get AI API key
  • Test API connections
  • Review security best practices

4. First run

Time: 20 minutes

Generate your first automated release notes and review results.

  • Run with sample data
  • Run with real repository
  • Review output quality
  • Identify improvements needed

5. Iterate prompts

Time: 30-45 minutes

Refine your categorization prompts based on results.

  • Analyze categorization errors
  • Update prompt with specific standards
  • Test improvements
  • Learn prompt engineering patterns

Prerequisites

Complete the Getting started guide before starting this tutorial.

  • Environment set up and tested
  • API keys configured
  • 30-60 minutes of focused time
  • Text editor open
  • Terminal or command line ready
  • Example repository in mind (or use provided samples)

Approach

This tutorial follows the plan-before-code principle. Each step discusses the approach before implementing, mirroring how you work with AI coding agents—ask for a plan first, then implement.

Use this tutorial

Option 1: Follow exactly (recommended for beginners)

Work through each step in order, completing all exercises.

Option 2: Adapt as you go (for experienced users)

Read through, adapt examples to your environment, skip what you already know.

Option 3: Reference guide

Jump to specific sections when building your own automation.

Get help

  1. Troubleshooting - Common issues and solutions
  2. FAQ - Frequently asked questions
  3. Examples - How others solved similar problems
  4. Your AI coding tool - Ask it to explain errors

Outcomes

After Step 2: Documented manual process—the foundation for automation

After Step 4: First automated release notes draft generated

After Step 5: Refined prompts producing high-quality categorization

Continue your work

After this tutorial:

  • Integrate into your release process
  • Try with different repositories
  • Adapt for different audiences (internal versus external)
  • Help others in the technical writing community

Time required

Total time: 2-3 hours

  • Initial setup: 30 minutes
  • Document process: 30 minutes
  • First runs: 30 minutes
  • Prompt iteration: 1-2 hours

This automation pattern can save 4-8 hours per release cycle.


Begin with Step 1: Setup environment