0.4.0

Added

• CLI bulk tagging for selected notes, folders, or a mix of both, with recursive folder selection, note exclusions, search, and sorting.
• Bulk dry runs showing selected notes, candidate tags, and estimated provider requests without API calls or note changes. Separate modes use Jev to return suggestions or apply matches, with up to three notes analyzed in parallel.
• Bulk progress, per-note outcomes, result filters, cancellation, and durable undo through a recovery file. Undo preserves notes edited after tagging.
• Added the option to provide a specific tag list for one note or a batch, including tags not yet in the vault, instead of using existing vault tags.
• Per-run CLI tag selection and matching overrides across single-note and bulk commands. Overrides leave saved settings unchanged.

Changed

• Renamed the single-note suggestion command to “Suggest tags for current note…” and the instant-apply action to “Match and add tags to current note.” Dialog headings follow the commands.
• Moved “Copy results” beside the results filter above the bulk results list.
• Tag-entry examples omit optional # prefixes.

Fixed

• Tagging dialogs previously jumped in height as content changed. They now keep a stable height and scroll longer content.
• Note titles were clipped at the top of the single-note suggestion dialog. They now display fully, including wrapped lines.

0.3.0

Added

• Bulk tagging for individual notes, folders, or mixed selections. Search and sort the file tree, then use Show selected to review the batch before applying tags.
• File-pane actions Match tags… and Match tags to selected notes… open the dialog with the selected notes already filtered. The command palette action is Match tags to multiple notes….
• Background tagging with up to three notes analyzed in parallel, reopenable progress and results, stop controls, and guarded undo that preserves later edits.
• Only these tags in global settings, individual reviews, and batch adjustments. Supply a specific set, including new tags, and match them independently against each note. This overrides vault selection and exclusions; match-score and addition limits still apply.
• The optional agent CLI supports the same explicit tag sets. Use Check CLI status in settings to check whether the installed companion needs updating.

Fixed

• Stable results-window height and middle truncation for long activity paths.
• Visible keyboard focus on folder disclosure and background-notice controls.
• Recovery instructions for provider errors and invalid note metadata.
• Single-note quick apply retains the settings used during analysis through the write.
• Invalid saved explicit tag lists can be corrected in settings without falling back to vault tags.

Validated with 112 automated tests, lint, TypeScript, plugin and CLI builds, packaged CLI smoke tests, and desktop Obsidian checks using synthetic notes.

0.2.1

Fixed

• Removed styles flagged by Obsidian's CSS checks. The review and settings look and behave the same.

0.2.0

Added

• Adjust the selection method, size, and mix for a single review without changing your defaults.
• Score more tags from the review results. Each round samples tags that haven't been scored yet and keeps your selections.
• Show more rows in long tag lists instead of stopping at 100.
• Check CLI status from settings.

Changed

• Redesigned the review to be easier to scan, before and after analysis.
• Search in review results also finds tags that weren't scored.
• You can add more tags than “Maximum tags to add” by checking them yourself. The limit still sets how many are preselected or added directly.
• The CLI update notice appears only when the installed CLI is out of date.
• Renamed the “Suggestions” settings group to “Recommended tags” and the pool dropdown to “Selection method”.
• Redesigned “Browse vault tags” in settings to match the review.

Removed

• The CLI evaluate command. The CLI now covers tagging only: preview, suggest, review, apply, and quick-apply.

Fixed

• The selection mix slider in settings no longer disappears after you change a selection setting.

0.1.6

Changed

• Shortened the Jev question repeated for each candidate tag.
• Made the CLI update notice optional. Plugin and CLI versions do not need to match after every release.
• Clarified per-tag scoring and companion CLI setup in the README.

0.1.5

Changed

• Clarified Community plugin installation and the controls for choosing tags in the README.
• Updated the plugin description to mention Jev and optional tagging guidance.

0.1.4

Changed

• The review dialog now evaluates the same tag selection and settings shown in its preview. It still reads the current local API key when analysis begins.
• The companion CLI accepts only the validated review-plan fields it needs when applying selected tags.
• Added focused review-flow coverage and two TypeScript lint checks for production code.

Verification

• CI passed lint, 52 tests, plugin and CLI builds, and CLI help.

0.1.3

Changed

• Separated the optional Node CLI source from the Obsidian plugin source. CLI installation and commands work as before.
• Use the active Obsidian window for retry timers, including popout windows.
• Clarified in the README how less-used tags can enter a limited selection.

Verification

• CI passed lint, 50 tests, the plugin and CLI builds, and CLI help.

0.1.2

Tag Match 0.1.2 makes API key setup across devices clearer.

• The TypeSafe and OpenRouter key settings state that API keys do not sync between devices. A saved secret name without a local key still shows a specific missing-key message.
• The README and agent guide explain how to set up a key on another device.

Download main.js, manifest.json, and styles.css below to install or update the plugin manually. The optional CLI assets are separate.

0.1.1

Tag Match 0.1.1 clarifies setup across devices.

• Settings shows when a saved Obsidian Secret has no key on the current device. Enter the TypeSafe or OpenRouter key in Obsidian Secrets on each device where you use Tag Match.
• The README explains what syncs and no longer presents a synthetic batching measurement as typical speed or cost. The benchmark scripts and results remain in the source repository for development.

Download main.js, manifest.json, and styles.css below to install or update the plugin manually. The optional CLI assets are separate.