Export a ChatGPT conversation to Markdown

Why this format

• Code fences keep their language

  Every block keeps its language identifier, so syntax highlighting works the moment the file lands in your editor.

• Tables stay tables

  Pipe tables are emitted properly aligned instead of collapsing into run-on paragraphs.

• Portable forever

  Plain text in a git repo, a notes app, or a static site generator — nothing proprietary and nothing to migrate later.

• Ready for coding tools

  Markdown transcripts paste cleanly into editors and coding assistants as reference context.

Example output

## Designing a typed conversion service

### Assistant · 14 Mar 2026, 14:02

Define the contract first, then write a mock that satisfies it.

```typescript
interface ConversationService {
  fetchConversation(shareUrl: string): Promise<NormalizedConversation>;
}
```

| Format | Best for | Fidelity |
| --- | --- | --- |
| Markdown | Coding tools | High |

Step by step

1. 01Paste the public share link and load the conversation.
2. 02Choose a heading depth so the transcript nests correctly inside your existing notes.
3. 03Enable YAML front matter if your notes system indexes metadata.
4. 04Download the .md file, or copy the whole transcript to the clipboard.

Tradeoffs and limits

• Inline formatting beyond bold, code, and links is simplified.
• Images are referenced by name rather than downloaded alongside the file.
• Very wide tables remain wide — Markdown has no column-width concept.

Files are generated in your browser. Nothing is stored.