An Obsidian plugin that automatically fills note frontmatter properties using configurable rules. Pairs well with Bases, Templates, and Dataview.
See the wiki for examples and full documentation.
You define rules in the plugin settings. Each rule watches note content and writes a value to a frontmatter property when its trigger fires. Note bodies are never modified — only frontmatter properties are written.
- Find in-work tasks
- Pull highlighted material
- Extract a TL;DR block
- Add the first image embed
- Do word counts or reading time
- Auto-add file create / update dates
- ...and whatever else you can think of
Pulls lines that match a pattern. Match modes: Starting with, Ending with, Containing, Regex.
Options: pull first / all / count of matching lines, omit the search string from the result, ignore indentation, pull the next line instead of the matched one, case sensitive.
Pulls content found between two delimiter strings (e.g. ==highlighted==, [[, ![[).
Options: separate start and end delimiters, retain delimiters in result, multiline matching, pull first / all / count of matches, case sensitive.
Pulls content from under a heading, targeted by level (1–6) or heading text.
Options: heading text only or full section content, include the heading line itself, include subheadings.
Pulls content from callouts, optionally filtered by callout type.
Options: extract header, body, or both; include type label.
Pulls file metadata: name, full path, folder, extension, created date, modified date, or file size.
Each rule independently controls when it runs:
| Trigger | When |
|---|---|
| Modify | Shortly after the note content changes |
| Open | Once when the note is first opened |
| Focus | When you navigate away from the note |
Rules with no triggers enabled are effectively manual-only — run them via the command palette (Auto-Properties: Update Auto-Properties).
| Option | Effect |
|---|---|
| Enabled | Toggle a rule on/off without deleting it |
| Auto-add | Creates the property in frontmatter if it doesn't exist yet |
| No overwrite | Skips the rule if the property already has a non-empty value |
These apply to the pulled value before it's written:
| Option | Effect |
|---|---|
| Pull | First match, all matches (as a list), or count of matches |
| Strip markdown | Removes markdown formatting from the result |
| Trim whitespace | Strips leading/trailing whitespace |
| Extract regex | Replaces the result with the first regex match, or the first capture group if present |
| Value format | Template string wrapping the result, e.g. https://example.com/${result} |
Available format variables: ${result}, ${filename}, ${folder}, ${path}, ${created}, ${modified}.
Each rule can be limited to run only in certain folders, or skipped in others:
- Run in folders — only process notes whose path starts with one of these
- Ignore folders — skip notes whose path starts with one of these
Ignore wins if both match. Leave both empty to run on all notes.
Rules can be exported as JSON to your clipboard and imported back (append to existing or replace all). This makes sharing rule sets or moving between vaults easy.
Example rules
[
{
"key": "overview",
"autoadd": true,
"strip_markdown": true,
"type": "callouts",
"callout_type": "tldr"
},
{
"key": "modified",
"autoadd": true,
"trigger": [
"focus_change"
],
"type": "file",
"file_pull": "modified"
},
{
"key": "task_count",
"autoadd": true,
"value": "- [ ]",
"pull": "count",
"ignore_indentation": true
},
{
"key": "highlights",
"autoadd": true,
"type": "between",
"pull": "all",
"start": "=="
},
{
"key": "img",
"autoadd": true,
"value": "![["
}
]- Note bodies are never changed. Only frontmatter properties are written.
- Everything optional except the property key. Sane defaults cover all omitted fields.
- No dependencies. Zero npm dependencies beyond the Obsidian API.
Listed in the Obsidian Community directory!
Or install via BRAT:
- Install & enable BRAT
- In BRAT settings, click Add Plugin
- Paste this repo's URL
- Done
Rules created in the earlier V1.0 will unfortunately be erased in the update to V2.0. You'll need to re-create your rules one time. Sorry for the annoying task. Shouldn't happen again.
This was a huge re-write and the rule parsing logic is much cleaner now. I don't expect to ever need to invalidate the current rule parsing enginge (this one is much more forward-thinking)!

