Text Transformation Prompt Library: reformatting dictated text with AI

I dictate a lot of text — probably more than I type at this point. Voice-to-text has gotten remarkably good, but the output is still fundamentally raw and unstructured: you get a wall of text that captures your thoughts accurately but looks nothing like a professional email, a set of meeting notes, a project proposal, or a to-do list. The gap between "what I said" and "what I need to send" requires a transformation step, and I was finding myself writing the same transformation prompts over and over. So I built the Text Transformation Prompt Library: a comprehensive collection of system prompts, each designed to transform raw text into a specific output format, available in both Markdown and JSON for easy integration into any workflow.

The dictation-to-deliverable pipeline

My typical workflow looks like this: I dictate a stream-of-consciousness brain dump about something — a project update, a client email, ideas for a blog post, notes from a meeting — and then pipe it through one of these prompts to get a polished output. The library covers a wide range of target formats: business communication (professional emails, meeting minutes, proposals, status updates), documentation (technical docs, READMEs, changelogs), content creation (blog outlines, social media posts, newsletters), task management (to-do lists, project plans, sprint items), format conversion (JSON, HTML, structured data), and style transformation (formal/casual tone shifts, clarity improvements, conciseness passes). Each prompt follows a standardised format with a name, description, the actual system prompt text, expected output format, and whether it delivers structured output. Need to turn rambling thoughts about a project into a structured proposal? There's a prompt for that. Want to convert meeting notes into action items? Covered. Need to draft a professional email from casual notes? Done.

Why a library instead of one flexible prompt

You might wonder why I maintain dozens of separate prompts rather than one general-purpose "format this text" prompt. The answer is quality: a prompt tuned specifically for converting meeting notes into action items produces significantly better output than a generic formatting prompt given the same instructions. Each prompt encodes domain-specific knowledge about what the target format should look like — the expected structure of a project proposal is fundamentally different from the expected structure of a technical README, and a purpose-built prompt captures those differences in ways that a general prompt can't. The library also includes an automation pipeline for integrating new prompts: you drop raw prompts into a to-integrate folder, and the automation handles standardising the format, generating JSON equivalents, and maintaining the index. The full library is on GitHub.