3.1.3

Fix issue with files names that include periods being recognized as attachments #653

3.1.2

Fix for #650

3.1.1

• Improved Notion API import speed by reusing page data, prefetching blocks, and adapting request pacing to avoid rate-limit stalls #646
• CSV reimports now preserve local edits unless the source CSV file is newer. #647
• Improved progress and completion button layouts on phones.
• Improved Apple Notes instructions on unsupported devices.
• Updated the YAML template filter to use the shared Knap implementation.

3.1.0

• Support for Templates across all importers.
• New Preview screen showing samples before import.
• Improved Android support for folders.
• Import from .txt files.
• Import from Logseq OG.
• Import from Bear's iOS Application Data export.

Templates

• Added a template and preview step across note importers. Importer generates a Markdown template automatically, which can be edited directly or replaced with a .md template from the vault.
• Templates support the same variables, filters, and logic as Obsidian Web Clipper.
• Added shared variables for note content, properties, title, destination path, source timestamps, importer ID, and source ID. Importer-specific variables are available for Airtable, Apple Notes, CSV, Google Keep, and HTML.
• Added live previews using examples from the selected source. Previews show the resulting note title, path, properties, and Markdown content.
• Template previews can be navigated, expanded, and edited using Obsidian’s native editor.
• Inline template edits are remembered for future imports. Note-title templates and managed source-ID properties can also be customized.
• Reorganized importer settings that affect note content or properties onto the template page.
• Normalize tags, aliases, and cssclasses generated by importers and templates to Obsidian’s supported list-property format (#609, PR #637, PR #643).

Logseq

• Added an importer for Logseq Markdown graphs. Choose a graph folder and select which page and journal folders to import.
• Convert pages, journals, namespaces, properties, aliases, tags, tasks, block references, embeds, queries, and linked assets.
• Read graph settings from logseq/config.edn, including custom page and journal directories, filename formats, and date formats.
• Journals can use the configured Obsidian Daily Notes folder and filename format.
• Added an optional Flatten outlines setting to convert Logseq outlines into Markdown paragraphs, headings, and lists.
• Scheduled and deadline task metadata is converted into linked dates.
• Queries can be preserved as fenced code blocks. Flashcard markers and time-tracking data can be retained as plain text.
• Preserve Markdown and Org code blocks, inline code, Unicode tags, nested paths, and source syntax that Obsidian cannot convert directly.
• Missing assets and unsupported whiteboards are reported instead of silently omitted (#47, PR #642).

Google Keep

• Import web annotations as a linked Markdown list.
• Added template variables for archived, pinned, and deleted state; color; labels; sharing data; and annotations.
• Labels are exposed to templates as an array of names.
• Add the Keep/Task tag when a note contains Google Tasks data.
• Expose source creation and modification times through the shared ctime and mtime template variables (#258, #385, #531, PR #639).

Bear

• Added support for importing Bear Application Data ZIP exports from iPhone and iPad.
• Preserve note creation and modification times, archived and trashed state, tags, and attachments.
• Improved Bear previews and embedded-image handling on mobile.
• Improved attachment matching across repeated exports and reporting for missing attachments. Encrypted notes are skipped and reported (PR #645).

Markdown and text files

• Added .txt support to the Markdown importer for individual files, folders, and ZIP archives.
• Text files preserve their source hierarchy and participate in relative-link repair like Markdown files (PR #644).

Other improvements

• Added native folder selection on Android, including recursive folder imports and source timestamps.
• Apple Notes templates can use {{isPinned}} to preserve whether a note was pinned (#156).
• Apple Notes titles derived from list items no longer retain the leading list marker.
• Roam imports now preserve unresolved block references and queries by default.
• Roam conversion no longer modifies markup inside inline or fenced code blocks.
• Improved responsive layouts for template navigation and Airtable property mappings.

3.0.0

Note this version requires Obsidian 1.13.0 or later.

• Importer is now available directly from Obsidian Settings. On phones and tablets, imports use native Settings pages with improved navigation, swipe gestures, and layout (#555 ,#634).
• Redesigned import steps with grouped settings, progress indicator, clearer status counts, and improved Pause and Stop controls.
• Imports left running in the background can be reopened from their progress notice. Starting another import waits for the current import to stop.
• Added searchable folder trees to make large source lists easier to navigate (#608).
• Improved file selection, especially on iOS. Selected files are listed individually, can be added or removed, and multiple selections are preserved.

Markdown

• Added a Markdown importer for individual files, folders, and ZIP archives (#23, #635).
• Preserve source folder hierarchies, relative links, attachments, .base files, and .canvas files.
• Added a folder tree for choosing which parts of a Markdown folder or ZIP archive to import.
• Added Standardize formatting to apply the vault’s list, indentation, and link preferences.
• Added Convert inline tags to properties to move tags from note bodies into frontmatter.
• Improved repeat imports by recognizing existing notes and attachments, avoiding unbounded numbered copies, and preserving locally changed attachments when Skip is selected.
• Fixed links when source paths must be sanitized, when folders differ only by letter case, and when source notes have names that differ only by case.
• Prevented imported attachments from overwriting unrelated files with the same name.

HTML

• Rebuilt HTML conversion around Defuddle to extract and standardize the main article content while removing navigation, sidebars, footers, and other page clutter. The complete page can still be imported by disabling Extract main content (#325, #636).
• Added support for importing complete HTML folders and ZIP archives, including linked local images and other assets.
• Added a folder tree for choosing which parts of an HTML folder or ZIP archive to import.
• Preserve HTML source folder structure and linked asset relationships (#406).
• Use the HTML document title for the imported note name instead of relying on the source filename (#325).
• Resolve links between imported pages, attachments, and headings, including same-page anchors and links to headings identified by HTML IDs (#217).
• Improved conversion of tables, figures, definition lists, fenced code blocks, Bootstrap alerts, GitHub pages, and React-generated markup.
• Preserve unusual HTML IDs and linked headings without allowing malformed CSS identifiers to interrupt conversion.
• Improved repeat imports by recognizing existing notes and avoiding accidental overwrites.
• HTML ZIP files are now recognized when dragged into the Importer.

Notion

• Removed the preliminary Notion page-count notice, which could be misleading before the final import plan was assembled.

2.6.2

• OneNote: More improvements for #630 issues with alt text.
• Escape pipes in tables #273

2.6.1

OneNote: sanitize alt tags. Closes #630

2.6.0

Apple Notes

• Remove redundant bold formatting when an entire heading is bold while preserving meaningful partial emphasis (PR #628).
• Ignore Apple Notes font family and point-size metadata, preventing unnecessary HTML while preserving bold, italic, underline, strike-through, superscript, subscript, and links.

Bear

• Keep tags inline by default. Added Move tags to a property to remove tags from note content and store them in frontmatter instead (PR #629).
• Improved tag detection around punctuation, escaped hashes, Unicode whitespace, code spans, code fences, numeric values, and hexadecimal colors.
• Convert Bear underlines to HTML <u> elements that work in Obsidian.
• Preserve image dimensions from Bear’s width and height metadata.
• Add blank line before Markdown tables.
• Create archive and trash folders only when the backup contains notes for them.

Roam

• Rebuilt Roam imports to convert outlines into natural Markdown paragraphs, headings, tasks, and lists. A new Flatten the outline setting can preserve the original outline instead (PR #627).
• Convert block references and embeds into working Obsidian links and embeds. Referenced blocks receive anchors, aliases are preserved, and missing references can be retained or removed (#246, #247).
• Convert top-level Roam attributes into properties and create a Base for browsing them. Attributes can optionally remain in the note (#245).
• Convert Roam tables into Markdown tables. Queries can be converted into Obsidian query blocks or removed (#180).
• Added settings for flattening outlines, embedding block references, removing missing references, retaining attributes, converting tags to links, and removing queries.
• Fixed links imported at the vault root, namespaced page links, daily-note links, long page names, and page names that collide after sanitization (#276).
• Improved headings, fenced code blocks, tasks, tags, style references, and video, audio, PDF, and iframe components.
• Roam attachments are now downloaded automatically, and existing copies are reused during repeat imports.

2.5.0

• Localized plugin across all 34 languages supported by Obsidian Help, including right-to-left language support (#607, PR #624).
• Added drag and drop. Drop files anywhere in the Importer window to automatically select the best matching importer, or choose between compatible import methods when a drop is ambiguous (#623, PR #625).
• Unrecognized files and folders can now be copied directly into the vault while preserving their folder structure. Textbundle packages and empty folders are handled correctly (PR #625).
• OneNote account sign-in now opens in the external browser, preventing authentication from stalling in Obsidian’s Web Viewer (commit).

2.4.0

New

• Offline Microsoft OneNote imports from .one and .onepkg files, without requiring a Microsoft account or internet connection. Imports support section selection, formatting, lists, tables, tasks, tags, highlights, equations, images, attachments, ink drawings, and internal links (#422, #580, PR #622).
• Import method selection for OneNote and Notion, separates account-based and local-file imports.
• Added automatic text-encoding detection from byte-order marks and declared XML or HTML charsets, including encodings such as GBK (#344).

Improved

Evernote

• Rebuilt the Evernote importer around Obsidian’s Vault APIs with streaming ENEX parsing, standardized duplicate handling, safer attachment placement, and clearer reporting (#403, #547, PR #620).
• Reimports preserve locally edited notes and established note locations, avoiding unnecessary rewrites, and reusing matching attachments.
• Links resolve against their final imported locations, recognizing Evernote app and web URLs, preserving unresolved links, and reporting unmatched links (#306).
• Preserve code-block languages, indentation, whitespace, and page indentation (#483).
• Preserve image dimensions, including different dimensions for repeated images (#49, #287).
• Preserve original attachment filenames without arbitrary truncation (#293).
• Improve timestamps, using the creation time when the updated time is missing and preserving modification times after link and attachment processing (#222, #286).
• Improve cancellation and pausing while retaining notes already converted.
• Removed unnecessary title sanitization.
• Replace legacy Evernote dot-numbered collision names with standard Obsidian-style suffixes.

Other

• Preserve OneNote section groups and complete nested subpage hierarchies (#546).
• Normalize imported filenames to Unicode NFC, preventing freezes and mismatches caused by decomposed filenames (#348).
• Allow Bear and Google Keep tags written in any Unicode script (#515).
• Improve Notion file-import error reporting. Unreadable nested ZIP archives now show the underlying error, and Markdown exports explain that an HTML export is required (#206, #364).
• Count only notes successfully written in the final “Imported” total; skipped and failed notes are no longer included. (#392)
• Improve Windows path-length calculations by accounting for the complete destination path.