Getting Started

This guide walks you through setting up Batch Generate Text with any AI and generating your first batch of content.

1. Install the app

Download and install Batch Generate Text with any AI from the Microsoft Store. Once installed, launch it from the Start menu.

2. Add your AI key

Before generating anything, add an API key from the provider you want to use.

  1. Click the key icon in the top-right header
  2. Select your provider from the dropdown - if you are not sure which one to use, see Choosing an AI Provider
  3. Confirm the model and Base URL
  4. Paste your API key into the field
  5. Click Save key
  6. Click Test connection to confirm the key, model, and Base URL work together

Your API key is encrypted and stored on your device only. It is sent directly to your chosen provider - nowhere else. You pay the provider directly; there is no subscription to Batch Generate Text.

Saving multiple connections

You can save up to 10 connections - handy if you use more than one provider, or keep separate keys for testing and production.

  • Each connection bundles a name, provider, Base URL, model, and API key
  • Use the connection dropdown in AI Provider Settings and choose Add new connection
  • Rename a connection to something clear, such as OpenAI test or DeepSeek longform
  • Select a connection to make it active - the provider, model, Base URL, and key all switch together
  • Delete connections you no longer need; at least one connection is always kept

Every key is encrypted on your device individually.

3. Choose a writing profile - Step 1

The persona controls the writing voice for the whole batch. It is a compact profile rather than a long prompt.

Starter personas included:

  • Chatty Betty - warm, conversational, relatable. Good for blogs and social posts.
  • Arty Marty - imaginative, vivid, expressive. Good for stories and creative pieces.
  • Serious Darius - precise, structured, professional. Good for analytical posts and explainers.

The persona view shows:

  • Writer type - the broad writing mode, such as brand voice, expert educator, storyteller, or analyst
  • Values - up to five brand or voice values
  • Goals - up to three output goals, such as inform, teach, persuade, relate, or entertain
  • Rules - evidence, CTA, edge, and rhetoric defaults
  • Do / Avoid - hard guidance the model should follow or avoid
  • Personal note - a short freeform note for nuance that dropdowns cannot capture
  • 10 writing datapoints - scored controls such as Formal/Casual, Serious/Funny, Neutral/Warm, Subtle/Direct, and Light/Dense

Click Edit to change the active persona. You can edit starter personas or add your own.

4. Add optional guidance - Step 2

Writing samples and quotes are optional. They help when you want the batch to match an existing voice or reuse approved source material.

Writing samples

  1. Click File to load one sample file, or Folder to scan a folder
  2. Folder imports are capped at 10 supported files
  3. Click Analyze to create a style report

Supported sample files include .txt, .md, .pdf, and .docx. The app can also load YAML sample banks where supported by the file picker.

If AI analysis is available, the report mirrors the persona structure: writing datapoints, sentence rhythm, paragraph pattern, vocabulary, openings, closings, rhetoric, do/don’t patterns, and persona-fit mismatches. If no API key is available, the app can still do a local scan and keep the sample text as raw guidance.

Quotes

You can load a quote bank if you want generated content to reuse approved source material. Quote files are YAML files with quote text, source, and optional theme fields. Matching themes help the app choose more relevant quotes.

Use the folder button in the app to open the built-in topics folder for example content lists, writing samples, and quote files.

5. Set up generation - Step 3

Step 3 is where you choose the output contract and run settings.

  1. Choose a Template - Blog Article is a good starting point
  2. Choose the Audience for this run
  3. Choose the Language near the top of Run Setup
  4. Choose Influence - Persona-led, Balanced, or Sample-led
  5. Choose the Output format - Auto uses the template default
  6. Select a Content list CSV or YAML file, or click Generate CSV
  7. Confirm the Output folder

Built-in templates

Template Example file
Blog Article blog-article.csv
LinkedIn Post linkedin-post.csv
X Post x-post.csv
Facebook Post facebook-post.csv
Short Story short-story.csv
Custom Template custom-content.csv

The app also includes a general articles.csv / articles.yaml starter list.

Generate CSV

The Generate CSV button creates a content list for you. Choose rows, type, topic, audience, mood, and theme; optionally add keywords and notes. Mood and theme can be multi-selected: moods are blended, while themes are distributed across generated rows.

6. Preview, then generate

  1. Click Preview to generate just the first item and review it in-app
  2. If the preview looks good, click Generate to run the whole list
  3. Watch the generation log for progress and errors
  4. Open the output folder when the run completes

Each run saves files in a timestamped subfolder, such as 2026-06-17_14-30-00, so runs do not overwrite each other.

Next steps

  • Try different templates to see which output format suits your workflow
  • Edit a template to adjust sections, output format, platform constraints, quote limits, and audience defaults
  • Add a quotes YAML file in Step 2 to weave curated quotes into generated content
  • Try Persona-led, Balanced, and Sample-led influence modes on the same first row to compare output
  • Use Generate CSV when you want a quick starter list without making a spreadsheet first