Getting Started with DevDiff

DevDiff turns your Git diffs into intelligent, human-readable changelogs — powered by AI that runs entirely on your machine.

---

Prerequisites

Before you begin, make sure you have:

Requirement	Version	Download
Node.js	20+	nodejs.org
Git	2.40+	git-scm.com
Ollama (recommended)	Latest	ollama.com

No Ollama? You can still use cloud AI providers like OpenAI or Anthropic. See the AI Providers guide.

---

Step 1 — Install DevDiff

npm install -g @eldrex/cli

Verify the installation:

devdiff --version
# DevDiff CLI  v1.0.3

---

Step 2 — Get a Local AI Model (one-time)

This downloads llama3.2:3b (~2 GB). Once downloaded, everything works offline — your code never leaves your machine.

ollama pull llama3.2:3b

Low on RAM? Use ollama pull llama3.2:1b (only ~1 GB). See model comparison.

---

Step 3 — Initialize Your Project

Navigate to any Git repository and run:

cd your-project
devdiff init

This creates a .devdiff.config.js in your project root. Done.

---

Step 4 — Generate Your First Changelog

Make a code change, stage it, and run DevDiff:

# Make a change
echo "// new feature" >> src/app.js

# Stage it
git add src/app.js

# Generate changelog
devdiff generate

You'll see output like:

📝 Changelog Generated

## Changes — June 30, 2026

### ✨ Added
- **src/app.js** — Added new feature entry point

### Impact: Low
### Files Changed: 1
### AI Model: llama3.2:3b (local, offline)

---

Step 5 — Try Different Personas

The same diff, explained differently for different audiences:

devdiff generate --persona developer    # Technical, code-focused
devdiff generate --persona ceo          # Business impact, executive summary
devdiff generate --persona compliance   # Regulatory, audit-ready language
devdiff generate --persona educator     # Teaching-style for junior devs

---

No AI? Use Dry Run

Preview what DevDiff would send to the AI — without making any API calls:

devdiff generate --dry-run

This works in CI environments with no AI configured.

---

What's Next?

• 🤖 Configure AI Providers — switch models, add OpenAI, or use WebGPU
• 🎭 Explore Personas — all 8 explained
• 💻 VS Code Extension — install the extension for inline annotations
• ⚙️ Configuration Reference — all config options
• ❌ Troubleshooting — Ollama not found? Start here