Train Your Voice

Train Your Voice lets you upload examples of your existing communications — past sermons, emails, blog posts, devotionals — so the AI learns your church's actual writing style instead of guessing from a tone description alone.

Where to find it

In the sidebar under Settings, click Train Your Voice.

How to upload a style source

1. Click Add Style Source
2. Paste the text content (or upload a document)
3. Give it a label (e.g., "Pastor's blog post on grace", "Easter 2025 email")
4. Save

You can add as many sources as you want. The AI references them all when generating content.

What to upload

The best style sources are pieces that sound like your church at its best:

• A favorite past email newsletter
• A pastor's blog post or sermon manuscript
• A bulletin you printed and were proud of
• A donor thank-you letter that hit the right tone
• An old social post that got great engagement

Aim for 3–5 strong examples rather than 50 mediocre ones.

What NOT to upload

• Templated form letters that don't sound like you
• Auto-generated content from another tool
• Anything you'd feel embarrassed showing a visitor

The AI learns from what you give it — garbage in, garbage out.

How it works under the hood

When you generate content, Midweek's AI is given your style sources alongside your Church Profile and weekly input. It uses them to match phrasing, sentence rhythm, vocabulary, and the small linguistic fingerprints that make your church sound like your church.

Tips

• Refresh sources occasionally. If your tone evolves (new pastor, brand refresh, demographic shift), swap in newer examples.
• Pair with Voice Profiles for distinct audiences. You can train the youth voice with youth-specific past content.
• Less is often more. 3 great examples beats 30 average ones.

Related: Voice Profiles · Church Profile