Guide

Markdown notes for AI workflows

Use plain, structured Markdown as readable input for AI tools while keeping private material under your control.

Markdown is a convenient way to prepare notes for AI tools because its structure remains visible as ordinary text. Headings indicate sections, lists group related points, and code fences mark text that should be treated literally. You do not need a special file format to make a note understandable.

The useful part is the separation between your source and an answer generated from it. A tool can consume Markdown, but the file remains yours to read, correct, or hand to a different process.

Make the structure explicit

Give the note a descriptive title. Use headings for distinct topics and lists for items you may want to extract later. Put exact instructions, queries, or logs in code fences. A little structure helps both human readers and software find the right part.

# Research notes

## Question
What changed in the onboarding flow?

## Evidence
- Interview notes from Tuesday
- Support themes from this week

## Open questions
- Which change should we test first?

Keep source and output separate

When an AI tool summarizes or transforms a note, keep the original note available. Mark generated material clearly and verify claims against the source before publishing or making decisions. Markdown makes it easy to compare a draft, a prompt, and a response as text, but the format itself does not make an answer correct.

A small, checkable workflow

  1. Write one note with a question, evidence, and unresolved points under separate headings.
  2. Remove private material that the destination tool should not receive. Share only the relevant excerpt if the whole file is unnecessary.
  3. Ask for a specific transformation: “Summarize the decisions in this note. Quote the heading that supports each decision. List anything the note does not answer.”
  4. Compare every cited passage with the original. Move useful output into a separate draft; keep your source note unchanged.

The last step is the important one. Markdown helps a model see structure; it does not grant it knowledge outside the text or guarantee that a summary is faithful. A clear “unknowns” section makes it easier to notice when a confident answer crosses that boundary.

Protect private material

Before sharing a note with any external service, inspect it for private names, credentials, customer data, or attachments. A local Markdown file is under your control until you paste, upload, or sync it somewhere else. The destination tool’s data practices matter more than the .md extension.

If you plan to move notes among tools, test exporting Markdown and preserve attachments separately. For syntax examples, use the Markdown guide.