ChatGPT to Obsidian YAML Frontmatter for Dataview and Bases

Exporting ChatGPT to Obsidian is most useful when the Markdown file includes structured metadata.

The note body is where the conversation lives. YAML frontmatter is what lets Obsidian sort, filter, and connect exported chats as a real knowledge base.

This guide explains how to use ChatGPT to Obsidian metadata with Obsidian properties, Dataview, and Bases.

Example Frontmatter

ChatGPT to Obsidian can export Markdown with fields like this:

---
title: "Competitive research summary"
date: 2026-07-06
originalDate: 2026-06-30
updatedDate: 2026-07-05
platform: ChatGPT
sourceUrl: https://chatgpt.com/c/...
sourceId: conversation-id
messageCount: 52
tags:
  - chatgpt
  - research
projectName: "AI Exporter Research"
---

These fields are intentionally simple. They work as plain Markdown, but Obsidian can also read them as properties.

What Each Field Is For

Field	Why it matters
title	Human-readable conversation title
date	Export date or note date
originalDate	When the ChatGPT conversation originally started
updatedDate	When ChatGPT last showed the conversation as updated
platform	Useful if your vault also stores Claude, Gemini, or Perplexity exports
sourceUrl	Link back to the original ChatGPT conversation
sourceId	Stable source identifier when available
messageCount	Helps identify long or high-context conversations
tags	Lets Obsidian group notes by topic or workflow
projectName	Keeps ChatGPT Projects searchable after export

Use Obsidian Properties

Obsidian will show frontmatter as properties at the top of the note.

Good property habits:

• Keep sourceUrl as a URL field.
• Keep date, originalDate, and updatedDate as dates.
• Keep tags as a list.
• Keep messageCount as a number.
• Keep projectName as plain text.

That gives you cleaner search and filtering later.

Dataview Examples

If you use the Dataview plugin, you can query exported ChatGPT notes.

List recent ChatGPT exports:

TABLE updatedDate, projectName, messageCount, sourceUrl
FROM "AI Chats/ChatGPT"
WHERE platform = "ChatGPT"
SORT updatedDate DESC

Find long conversations:

TABLE messageCount, projectName, sourceUrl
FROM "AI Chats/ChatGPT"
WHERE messageCount > 40
SORT messageCount DESC

Find conversations from one project:

TABLE updatedDate, messageCount, sourceUrl
FROM "AI Chats/ChatGPT"
WHERE projectName = "Website Refresh"
SORT updatedDate DESC

Find Deep Research notes:

TABLE updatedDate, sourceUrl
FROM "AI Chats/ChatGPT"
WHERE contains(tags, "deep-research")
SORT updatedDate DESC

Obsidian Bases Ideas

If you use Obsidian Bases, the same frontmatter can power table-style views.

Useful views include:

• All ChatGPT exports sorted by updatedDate
• Project conversations grouped by projectName
• Long conversations filtered by messageCount
• Deep Research reports filtered by tag
• Archived Chats filtered by folder path or tag

Suggested Tags

Do not over-tag every export. Use a small controlled set:

tags:
  - chatgpt
  - deep-research
  - project
  - coding
  - writing

If a note belongs to a ChatGPT Project, use projectName for the project and keep tags for broader categories.

Best Practices

Keep Source URLs

The Markdown file is your local copy, but sourceUrl helps you return to ChatGPT when the original context still exists.

Use Updated Dates for Re-exporting

If a conversation changes after the first export, updatedDate makes that visible.

Keep File Names Stable

For active notes, use updated-date naming:

{updated_date} - {title}.md

For archival notes, use original-date naming:

{original_date} - {title}.md

Do Not Put Everything in Tags

Use fields for structured data. Use tags for broad categories. This keeps Dataview and Bases cleaner.

Final Thought

YAML frontmatter is the difference between a folder of exported chats and a usable Obsidian knowledge system.

Keep the metadata from ChatGPT to Obsidian, then use Obsidian properties, Dataview, or Bases to turn exported conversations into a searchable archive.