AI Chat Connectors 2.0

When we launched AI Chat Connectors, your connected assistant could search your space, read full objects, save text to your daily note, and link to objects in its replies.

This release adds new capabilities to all AI Chat Connectors and MCPs:

• Create pages and tasks: You can now instruct your assistant to create pages and tasks in Capacities.
• Create custom objects: By mentioning the name of an object type (e.g. "Create a meeting"), your assistant can create a new object of that type with the properties you specify.
• Update object properties: You can instruct the AI to change properties on existing objects (status, tags, dates, title, and custom properties).
• Append content to objects: You can instruct the AI to append content to existing notes.

Together with the existing read tools and the ability to save text to your daily note, you can run fuller workflows from the AI apps you already use.

Example prompts

• "Create a Capacities page titled 'Weekly review template' with my checklist as markdown."
• "Add a high-priority Capacities task to finish the connector docs by Friday."
• "Set the status on my Capacities Book object for 'Deep Work' to finished."
• "Append a summary of this conversation to my meeting note about the product launch in Capacities."

Setup

Nothing changes in how you connect a tool: use the same connection URL and flow as before:

• AI Chat Connectors documentation
• MCP tool reference

You need Capacities Pro with AI enabled for the space you connect.

Note: To use the new capabilities, you may need to reconnect your existing tools. All connections can be managed in Settings → AI Assistant.

Roadmap

AI Chat Connectors will continue to evolve alongside our upcoming new public API. 👀 We'll be developing and improving both in tandem, adding more powerful capabilities over time.

That said, if you're building workflows that involve a lot of content interaction, we recommend using the AI assistant in Capacities. It offers more powerful retrieval and search features already today, and will also receive more advanced editing and creation capabilities in the future.

We've also made a video version of our release notes to give you a quick overview.

✨ Improvements - A Big Step Up in Quality

Since the last release, we have shipped a wave of polish across the app: smoother tabs, a more comfortable editor, better drag and drop, and fixes in areas you use every day.

Object dashboard and tasks

• Query suggestions in the object dashboard: When you add a new tab on an object type overview, Recommendations offer ready-made views (filter by a label, show only checked or unchecked items, or a kanban grouped by labels).
• Tags view for the task dashboard: By default, you can now view all tasks grouped by the tags they are assigned to.
• You can now edit queries and collection titles from the object dashboard.
• Task context picker: The context picker for tasks is now standardized. You can select multiple options, search more effectively (including exact matches), create new contexts when needed, and create a page directly from the menu.
• Label editing: If a label cannot be edited directly, you will now be guided to the settings page where changes can be made.
• Added a small card preview when hovering over multi-select properties, making it easier to see details quickly.

Tabs

• Space tabs and side panel tabs now look the same way.
• When you have many tabs open, they shrink to fit instead of scrolling off the screen.
• An overflow menu appears when tabs get too cramped.
• You can pin and unpin tabs from the right-click menu on tabs.
• Tab switching feels smoother, with fewer visual glitches.
• You can close a sidebar tab by double-clicking it, speeding up tab management.
• Object preview when hovering over tabs gives you a quick overview without opening the object.

Editor spacing and layout

• Text blocks have improved spacing and rhythm throughout, making longer notes feel more comfortable to read and write.
• Headings now follow a cleaner size scale.
• Paragraphs have a bit more breathing room.
• Lists, to-dos, and toggles have clearer spacing and a nicer selection highlight when mixed together.
• Linked objects (entity blocks) have more balanced spacing above and below.
• Horizontal dividers have more even spacing, especially when followed by lists, to-dos, or linked objects.
• Table cells have rounded corners and a clearer focused state.
• Two-column blocks are hidden from the create menu on phones, where they aren't supported.

Block linking

• The reference counter no longer sits on top of your content. It now flows beside the block instead.
• The reference menu opens as a proper bottom sheet on mobile.

AI assistant

• The AI Chat now has a stop button whenever a response is being generated.
• You can now have multiple AI Chats working concurrently in the right side panel.
• Overall improvements in speed and error handling for all AI features.
• The AI Assistant can now read specific sections of analyzed media and distinguish your notes from media transcripts.
• AI replies now use the full width of the chat column, giving activity indicators and source pills more room to breathe.
• Sources are now shown as a cleaner grouped card with smaller text and website names, making references easier to scan.
• The context bar above the chat input is hidden for now. Use @-mentions to add context instead.

Drag and drop

• You can drag an inactive tab to the right side of Capacities to open the side panel automatically when it is hidden.
• You can drag objects from various views and the space overview (the left sidebar) into the side panel.
• You can drop individual objects into collections and tags views, where applicable.
• Drag and drop now supports dropping onto custom space overview sections, cover images, and specific objects.
• Nested drag-and-drop areas are handled more smoothly.

Images and links

• Broken image tooltips look cleaner and easier to read.
• Web link preview cards have a tidier cover image layout.

Import, sync, and files

• Property synchronization during imports now preserves two-way links correctly.
• Added an option to always allow when opening local file links (file://) to simplify access.
• Added import action buttons in empty states of media views to make importing easier.

Onboarding

• The "Tip of the Day" cards have more breathing room and a cleaner look.

🐛 Fixes

AI and chat

• AI panel is working again after the last release.
• Disabled autofill instructions in spaces that are excluded from AI features to avoid confusion.
• Clicking the "Tasks" tab in an AI chat now correctly toggles the content display.
• Corrected AI usage when preferring external API keys in the AI Assistant settings. If you were seeing budget errors with this option enabled, they should now be resolved.
• AI Chat error "model not found" should now be fixed for all providers.

Editor and blocks

• Indenting a heading no longer resets it to normal text. Your H1–H4 styling is kept.
• Using backspace with the cursor at the start of a text block now first unindents and only merges with the block above when the block can't be unindented anymore.
• Copying a block reference now works on mobile, just like on desktop.
• Inline code formatting (⌘E) now toggles off reliably and keeps the cursor in the right place after applying it.
• Block references can now be selected with a rectangle selection.
• Pressing Backspace or Tab near the edge of a block reference no longer accidentally moves text into the referenced item.
• All block references are now counted correctly, including in code blocks.
• Indented child blocks under text blocks stack correctly again after the layout update.
• Fixed a layout issue where reference arrows would cover text, so text is no longer hidden behind these arrows.
• Changing the type of an embedded note no longer changes how the note is displayed.
• Fixed an issue where presentation mode would lose your place when stepping between blocks.

Data views and properties

• The calendar button on date fields opens the date picker again as expected.
• Fixed the expand button on date properties so the calendar opens correctly again.
• Embeds in data views expand correctly again.
• The width of the drag-and-drop activation area in Kanban views has been reduced to prevent accidental interference with other interface elements.
• Fixed inconsistent display in the query editor and cleaned up selection menus for clearer data views.
• Resolved an issue where slider icons disappeared when grouping items in table view.
• Label values now display correctly on small cards in gallery view.

Images, gallery, and exports

• Fixed spacing in the image gallery to improve how images display.
• Restored cover image actions (expand, edit). Edit options work reliably again when dragging items over them.
• Highlighted text and cover images now render correctly in PDF exports.

Side panel and tabs

• Embeds in the side panel work again, even when the same page is open in the main view.
• The close button on side panel tabs is no longer clipped or cut off.
• Object previews are disabled when hovering over the active space tab to avoid unnecessary distractions.

Sidebar and navigation

• Sidebar item reordering works again after a drag-and-drop issue was fixed.
• Fixed an issue with the "Move To" action that prevented it from working correctly.
• Fixed filtering in backlinks so it now works correctly.
• Unlinked mentions only show up when there is also a backlink to the note, preventing confusion.
• Fixed shift+click on the calendar's Today button to open daily notes in the side panel again.
• Resolved problems with nested drag-and-drop areas to prevent conflicts and improve usability.

Daily notes

• Tags are now correctly shown on the daily note page again.
• The template menu (three-dots icon) in the daily note header is now always visible, even before a note exists, so you can create a note from a template without any extra steps.

Mobile

• The linking menu stays visible above the keyboard instead of being hidden behind it.
• Notifications and feedback messages now stay visible on top of open sheets and modals, instead of being hidden behind them.

Import and settings

• Fixed an issue where icons did not display correctly in the import menu.
• Corrected the setting to hide structure options so it only applies to new users.
• Renamed the "Upgrade" label in security settings for clarity.

Search and performance

• Several search issues reported in nightly have been fixed.
• Removed some heavy background diagnostics that were causing noticeable UI lag for some users.
• Resolved a couple of performance issues after the latest nightly update.

Original source

This release brings Bulk Import for everyone, Image Analysis for Pro, and a round of continuous improvements across tasks, AI chat, and everyday polish. 🚀

Bulk Import for Everyone

All users can now bulk import Markdown, CSV, ZIP, and media files!

We recommend a three-step process to get the most out of bulk import:

1. Review & sanitize — clean up your source files before import.
2. Object types — set up the object types you will need.
3. Importer — run the import and review objects before creating them in Capacities.

For setup instructions, supported formats, and limits, see the Bulk Import reference.

Thank you to all Believers who tested Bulk Import and gave us feedback! 🙏 It won’t be long until the next feature is available for testing! 👀

Image Analysis for Capacities Pro

AI image analysis is now available for all Capacities Pro users with AI enabled.

Open any image and click Analyze to extract a title, description, text (OCR), color palette, and category. Everything is written to the image's properties automatically.

We've also added several improvements since the initial release:

• Analyze from the Object menu now uses the same logic and budget checks as the toolbar action.
• Clearer AI budget feedback: if your budget is exhausted, the action shows "AI budget used," highlights with a yellow sparkle, opens AI settings, and surfaces clearer errors.
• Image embed affordance now includes a quick Analyze trigger plus OCR/colors preview, with a scan icon when analysis is available.
• New output categories: Photo (animal) and Cover.
• Localized labels: analysis labels now follow your app language.
• Cleaner image analysis UI: the separate analysis tab is removed; extracted text and color swatches are shown directly in the image toolbox and on the image.

If you want the full walkthrough, see the Images documentation.

Continuous Improvements

We have also shipped quality-of-life updates across the app. 🚀

Recurring Tasks Updates

Building on Release 64, recurring tasks now handle more scheduling patterns.

Monthly recurring tasks can repeat on the last day or last weekday of the month

In Custom repeat, when you choose a monthly schedule, you get two new "On" options alongside the existing day-of-month and Nth-weekday choices:

• Last day — the true end of each month, so February and 30-day months are handled correctly instead of skipping when you used the 31st.
• Last {weekday} — e.g. last Wednesday, which works with any interval such as every 2 or 3 months.

That covers monthly reviews, payroll-style tasks, and other "last X of the month" patterns without workarounds.

Recurring tasks now carry their deadline forward automatically

When you complete a recurring task, its deadline advances to the same relative position on the next occurrence. So if your task is scheduled on the 6th with a deadline on the 7th, completing it moves both dates to the following month together.

If you were behind and the recurrence had already jumped ahead, the deadline advances too and is clamped to at least the new scheduled date, so it never shows as missed the moment it is set.

On top of that, the calendar's "today" view no longer surfaces recurring tasks whose next occurrence is still in the future just because an old deadline has passed. Those tasks correctly appear on the day they are actually scheduled.

Reworked Sidepanel

We reworked the side panel to make it easier to discover and use. Instead of showing a separate button for every side panel action, there is now a new Explore tab that opens automatically and also acts as the new blank side panel tab.

From Explore, you can open relevant content or contextual options such as graph view, backlinks, objects inside, and related content. Each option now includes a short description, so it should be clearer what it does before opening it.

This should help declutter the side panel while keeping all options accessible.

We also deprecated the "Swap" button, replacing it with drag and drop. You can drag your side panel tab to the main window, and it will swap automatically.

More standardized indentation logic in text editing

We fixed the indentation behavior for lists and other blocks to match industry-standard editor behavior. Outdenting no longer pushes items into a different block position, so bullet order stays stable instead of being reshuffled. This applies both when editing inside text and when indenting/outdenting selected blocks. If one block in a multi-selection cannot be indented, none of the selected blocks move.

AI Web Search

The AI Assistant can now search the web to answer your questions with up-to-date information.

When you ask something that benefits from current or external information, the assistant will automatically run a web search and incorporate the results into its response. This works most reliably with Capacities AI — custom model configurations may have limited or no web search support depending on the provider.

AI Chat Improvements

The AI Assistant has had many improvements in the background. It should now be able to break down more complex tasks and give better responses in general. Some extra changes:

• The AI Assistant now also has access to our support documentation.
• Tool pills polish: gerund labels while a tool is running ("Searching…") and past tense once finished ("Searched"), plus richer tooltips built from the tool's inputs, outputs, and resolved titles.
• The "Thinking" pill is restored during the gap between submission and the start of the assistant stream.

Community Integrations

Two new community-built integrations are available in our Integrations Hub:

Community integrations are not part of the official Capacities product. We review and test integrations and work with developers on quality, but we cannot guarantee them or be responsible for issues from using them.

MelliLex

MelliLex is a free Windows desktop app that lets you Ctrl+right-click any word or phrase in any application to get an instant AI-powered breakdown, then save it directly to Capacities as a structured word note.

The Capacities export creates a word note with the looked-up term as the title and fills in the full lookup result — definition, IPA, part of speech, formality score, usage domains, example sentences, and memory tip. You can send notes to your daily note or a dedicated Vocabulary space.

Maintained by Rahul TR. Download free on GitHub.

CapNote

CapNote is a macOS menu-bar app that opens with a keyboard shortcut and appends notes directly to your Capacities daily note.

Maintained by Leandro Ardissone. Learn more on the CapNote website or GitHub.

✨ Improvements

• Want to try Capacities Pro again? If your last trial ended more than 6 months ago, you can now start another free trial.
• AI-enabled behavior for new spaces: If you have AI enabled but only for some spaces, it will now be enabled for new spaces automatically.
• Better search ranking in linking menus: We now show the create option first in the link and create menus if there is no direct object match.
• If you press Esc in a task title, Capacities will stop auto-filling properties based on keywords like "today."
• Improved background sync behavior for better continuity.
• Improved the alias menu and mobile experience for shortcuts and tag handling, fixing several glitches with menu placement and state.

🐛 Fixes

• Raycast Integration is fixed again. You might need to create a new connection in the Raycast settings in Capacities.
• Some layout fixes and issues with property sections.
• Addressed issues that could cause side panel behavior to feel inconsistent.
• Improved overall stability around data processing and error handling.
• Resolved several edge cases that could lead to stalls or interruptions.
• Fixed an issue where the object menu button was incorrectly shown on mobile views.
• Fixed the issue preventing moving the Capacities window by dragging on empty space of the tab bar.
• Corrected double window drag issues on desktop for easier repositioning.
• Opening the views in the side panel now works as expected again.
• Fixed the floating sidebar in daily notes so it now correctly scrolls to headings.
• Fixed task progression on mobile devices to display correctly without having to open the task.
• Fixed missing view options in the tasks tab for objects.
• Fixed UI issues by hiding certain elements when the user does not have edit permission.
• When you create an object from the day view, the date field is pre-filled correctly (fixes missing date on create-from-day).
• Small card view settings for basic object types are saved reliably again after customization (no more lost layout when reopening settings).
• Re-applying media analysis properties now only happens on explicit retries: there was a bug where media analysis results constantly overwrote properties, for example when they were edited by the user. This is now fixed.
• Mobile image analysis toolbox and OCR badge behave more reliably on small screens.
• Added text displaying the percentage of AI budget usage to the AI Assistant settings.
• Added missing autofill options to the customize menu to make all properties available for easier setup.
• Fixed the task status shortcut so it now includes custom statuses correctly.
• Corrected the default view for tasks so it displays as expected every time.
• Improved the Mac app search so shortcut keys should now work reliably. Please let us know if you still experience issues.
• Fixed the order of task groups in the relative date view so "This week" appears in the correct place.
• Fixed an issue in the Mac app that made the window draggable in the wrong place when adding additional areas to the side panel.

From Us

✍️ Subscribe to our Substack

There are lots of pieces on there now!

• Note-taking is synonymous with presence
• Note-taking is meant to be messy
• What do shipping containers have to do with note-taking?
• Rhizomatic note-taking starter pack

We're sharing deeper product thoughts, feature context, and building updates in a longer-form format there.

📺 Follow Capacities on social media to see our recent content on the team's favorite features:

• Twitter/X
• Instagram
• YouTube
• LinkedIn
• Reddit
• Threads
• Bluesky
• Mastodon

That's everything for this release – we hope you like it! ✅ 🚀

Original source

Check out the video release notes!

Recurring Tasks

This is one of our most requested features: Capacities now has full support for recurring tasks out of the box! 🚀 ✅

You can now set a task to repeat and keep working from the same task object. When you complete it (or skip it), Capacities logs that occurrence and moves the date to the next one automatically.

Recurring tasks support both ways people actually plan:

• Scheduled Date for fixed rhythms like every Monday or every month on the 15th
• Completion Date for flexible intervals like every 3 days after you finish

You can also customize the recurrence patterns in detail, add an end date, and much more. You can explore it in the dropdown in the date picker.

Capacities helps you when working with tasks. If you missed a recurring task you can simply skip it. Also if you know you can't do a task (perhaps you're on holiday), you can excuse the task and keep your streak.

If a recurring task is overdue, you have the option to also move it to the current or next occurrence.

Task Stats (Experimental)

Each recurring task includes an occurrence log, streak stats, and a heatmap so you can see how things are going over time.

On top, you can see a Completion log of your recurring task below.

All completed recurring tasks show up in your "Completed" section of the tasks dashboard.

You can all get a bird's eye view of all recurring tasks in the new 'Recurring' section in the task dashboard.

All future occurrences of a task are rendered with a dashed line around the check box in your calendar.

If you want to know all about recurring tasks, you can read the documentation.

Choose your model provider for Capacities AI

Capacities AI now supports multiple LLM providers (OpenAI, Gemini, Anthropic, Mistral and xAI). You can choose your preferred model in the AI chat every time you start a chat.

The model picker within the AI Assistant shows more information about the model.

• Capacities host as many models as possible in European data centers under strict data protection laws. You can now see which models are hosted in the EU. Learn more.
• Usage cost. You can see how quickly the model uses your AI budget.

In the settings, you can also choose your preferred model provider. Capacities will use it as a default for all AI features that allow the selection of a model. Learn more.

Improvements & Fixes

✨ Improvements

• Restored the option to customize and hide details in query headers for better data views control.
• Added /standard and /small commands to the create menu for quicker access to these block types.
• Made the customize button show a highlight effect when hovering over it for easier discovery.
• Improved the visibility of progress rings of tasks so you can better see task progress.

🐛 Fixes

• Fixed page view to show all tags and collections as before for clearer object displays.
• Fixed the object menu on mobile so adding cover images works properly on Android devices.
• Increased the allowed number of objects during bulk import for smoother data handling.
• Improved the behavior of back buttons in mobile sheets to make navigation more reliable.
• Fixed app and site freezing issues in object type settings on mobile devices.
• Prevented the deletion of objects when pressing enter in an empty embed title.
• Made the general entity picker snap to fullscreen on mobile to ensure all options are visible.
• Restored the customize object submenu in the object menu on mobile.
• Enabled back navigation in popover submenus on mobile for easier menu navigation.
• Fixed wrong wording in Italian on the task page.
• Corrected backlink count showing as zero on tasks query.
• Fixed markdown export not downloading properly.
• Corrected the shortcut shown in the shortcuts window for inserting icons to match the subscription state.
• Fixed the replace button not working after AI text actions.
• Fixed a freeze issue when loading the AI panel model.

From Us

✍️ Subscribe to our Substack

We're sharing deeper product thoughts, feature context, and building updates in a longer-form format there. First article already out!

📺 Wesley Anna's latest Capacities video

📱 Follow Capacities on social media to see our recent content on the team's favorite features:

• Twitter/X
• Instagram
• YouTube
• LinkedIn
• Reddit
• Threads
• Bluesky
• Mastodon

That's everything for this release – we hope you like it! ✅ 🚀

Original source

Release 63

Release 63 helps you find what matters faster, plan tasks more clearly, and keep AI chats focused while you work.

In this release:

• Related Content on every object for Capacities Pro
• Task management updates with deadlines, smarter task views, and progress rings
• Search 3.0 with more relevant results and query modes
• AI Assistant upgrades with better context control and chat usability
• UI upgrades across page views, side panel, and app shell
• Full Bulk Import access for Believers

Check out the video release notes!

Related Content

A common question for knowledge workers is how to resurface content at the right time.

Linking and tagging already help with this, but Related Content is another way: Capacities resurfaces related notes based on the object you currently have open.

You get the closest 5 related objects in a section directly below your notes, so discovery happens right where you are already working.

You can open Related Content in the side panel and scroll through more results without leaving your current context.

Related content works with AI Image analysis too, meaning it can resurface relevant images too.

Related Content is turned on by default and is available for Capacities Pro and Believer users. If you prefer not to see it, you can change this in Settings > Editor and hide it or keep it closed by default.

Task updates

Deadlines for Tasks

Tasks now support deadlines as an optional property.

You can set a deadline when creating a task or directly on the task page. Capacities also shows how much time is left, so it is easier to prioritize and plan your work.

You can also review all tasks with deadlines directly in your task overview.

This new feature is already integrated into the task views.

• Inbox: tasks with only a deadline do not show up in Inbox (they count as triaged).
• Today: overdue and due-today deadlines appear alongside overdue/today scheduled tasks.
• Future day view: each day now shows tasks scheduled on that day or due on that day.

Across the main task views, tasks with deadlines are ranked above tasks without deadlines. Earlier deadlines come first, used as a tie-breaker after priority and scheduled date.

Progress ring for tasks

For tasks that contain to-dos or tasks inside, the checkbox now shows a subtle progress ring so you can see completion at a glance.

Task creation and planning improvements

• You can now mention objects and tags in task titles by typing @ or # when creating tasks
• Better behavior of the task Context dropdown.
• You can drag and drop tasks in the Scheduled kanban view to move them to another day.

Search 3.0

Search now does a better job reflecting your intention, so you can find what you are looking for even when you do not remember the exact wording.

Results now include conceptually related content, not just exact phrase matches.

Fallback behavior is less fuzzy and more useful, preferring relevant results over weak partial matches.

Search queries and extended search now support modes (Auto, Exact, Semantic) so you can control how results are returned.

Breaking change: The default mode of search queries changed to Auto. If you want to restore the previous behavior as best as possible, we recommend switching to Exact.

The new search is available in the command palette, queries, extended search, object dashboards, collections, and tag pages. If you link to objects and select an object type, linking search also uses the new search algorithm. This improves discovery in both everyday search and larger knowledge bases where wording varies a lot.

AI Assistant upgrades

You can now keep context precise in AI chats without restarting your conversation, simply link to whatever objects you need with @ in any message in your chats.

This makes follow-up questions more useful, reduces context noise, and helps you get better answers without opening a new chat.

We also improved the day-to-day chat experience:

• Better scrolling behavior for incoming replies.
• Chats auto-save after the first message.
• Automatic chat title generation after the first assistant reply.
• Improved loading states with smoother visual feedback for tool calls.
• General UI improvements.

UI upgrade

This release includes a broad design improvement across the app:

• Refined page layout for a more unified and balanced page view.
• Improved side panel behavior and moved side panel tabs to the top level.
• Added an optional wide layout for daily notes.
• Polished shortcuts menu and snackbar/toast UX.
• Added hover preview cards in helpful cases.
• Simplified space overview actions/sections for a cleaner look.
• Reworked previews and export rendering (including AI chat and PDF output) with cleaner typography, spacing, tables, code blocks, and Mermaid rendering.

Full Bulk Import for Believers

Bulk Import is now fully available for Capacities Believers.

We also shipped several reliability improvements based on feedback from the initial launch. Thank you to everyone who tested and shared feedback! 🙏

This release also improves bulk import behavior in key edge cases:

• Relative file paths without ./ are now recognized.
• Full-line Markdown embeds are converted into separate object blocks.
• Creating a new custom type during import no longer creates an empty object automatically.

For setup instructions, supported formats, and limits, see the Bulk Import reference.

We'll do another round of feedback and improvements, then we will release bulk import for everyone. 🚀

What’s Next?

We recently updated our What’s Next article so you know what we’re focusing on in the coming months.

Thank you so much for your support. It means a lot and helps us keep building. 🙏

✨ Improvements

• We moved the notes into the media analysis tab, so you can view media analysis while seeing and working in your notes.
• To edit sections in your object dashboards created with queries, click on the section name and click 'Open'.

🐛 Fixes

• Cmd + A no longer selects Related Content and Mentions.
• Fixed inconsistent title focus behavior in task embeds.
• Task embeds no longer overflow on smaller screens.
• Fixed bulk import support for importing multi-line properties into multi-select fields.
• Fixed bulk import handling of multiple values for multi-select fields.
• Fixed bulk import issues with CSV files.

Original source

AI Chat Connectors

Your Capacities space can now talk to the AI tools you already use. 🚀

With AI Chat Connectors, Capacities Pro users can connect their space to ChatGPT, Claude, Le Chat, Cursor, and more using the open Model Context Protocol (MCP) standard.

Once connected, your AI assistant can search your notes, read full objects, add content to your daily note, and even link objects directly in its responses.

Ask ChatGPT what you've written about a topic, have Claude store a summary in your daily note, or let Cursor pull in context from your space while you code.

Each connection is scoped to a single space, so you stay in control of what your AI tools can access.

Set it up in just a few clicks from your AI tool or from within Settings > AI Assistant > AI Chat connectors. For full details, check out the AI Chat Connectors documentation.

AI Chat Connectors are currently in Beta, and we'll continue expanding their functionality over time. You need to be a Capacities Pro user with AI enabled to make use of this.

Bulk Import (Beta) for Capacities Believers

We're excited to introduce a Beta version of Bulk Import for all Capacities Believers.

You can now import up to 100 objects at once, making it much faster to bring existing content into your space. This is the perfect time to import smaller CSVs, multiple images, or a few selected notes to enhance your existing Capacities space.

We recommend a three step process to get the most out of bulk import:

1. Review & Sanitize - Review and clean your data before import.
2. Object type - Create the object types you will need.
3. Importer - Run the import and create your objects in Capacities.

This is the initial testing phase of Bulk Import. We'll expand limits based on your feedback. Because bugs might still occur, we understand if waiting for the full Bulk Import release is the better option for you.

After this initial testing period, we will open Bulk Import to Believers with no restrictions. After that, we will launch it for everyone.

You can read the instructions here or see it in action in this video.

✨ Improvements

• Better storing infrastructure: We reworked how storing your content works. This should especially improve high-load situations. You should not notice a difference, but if you're having issues with sync and persistence, please let us know.
• The “Customize” option in the object menu is now displayed on all devices, including desktop.
• The formatting of file sizes is now limited to two decimal places.
• Added support for matching properties in unlinked mentions to make finding related information easier.
• Improved filtering options in object picker modal to better target specific objects.
• After saving an AI chat as an object, the three-dot object menu now appears as expected.
• Properly escape type and collections property values when exporting to Markdown.

🐛 Fixes

• Aliases (if enabled in the export menu under properties) can now be included in PDF exports.
• In Kanban and Wall view, the Created at and Last Updated properties now show up as expected when enabled for Small Card View in Object Type Settings.
• Fixed an issue that caused Markdown imports into an object to fail in certain scenarios on Windows.
• Formatting issues that caused export errors and blocked exporting certain objects have been fixed.
• Full Export is now more robust and can complete even if errors occur during export of some objects.
• Disabled the option to export shared objects.
• Fixed an incorrect declension in the “Create Object” button in German.
• It is now possible again to customize small card views of basic object types via settings.
• The content outline is now only displayed in the “Content” tab of an object, no longer in the “Task” tab.
• Fixed translation issues in the date picker for some languages.
• Fixed linking issues when inserting objects by alias.
• Corrected template handling so meetings created from calendar events apply the right templates.
• Fixed export of date-time properties so both start and end times are included in Markdown exports.
• The task inbox now shows a clear empty state message when there are no tasks.
• Fixed search issues in the linking menu so results display correctly.
• Fixed the Table of Contents indicator so it correctly follows page position while scrolling.
• Resolved an issue with extended search where opening and scrolling in data views did not work as expected.
• Fixed onboarding overlay issues on Android devices.
• YouTube embeds now work as expected again.

Original source

Remember the new image properties we added in the last release? Now the AI can fill them in for you. ✨
Media analysis is available in Beta for all Capacities Believers with AI enabled. It turns your screenshots, documents, and photos into information you can easily search, filter, and work with. 🖼️
Check out the video release notes!

Analyze any image with one click

Open any image in Capacities and click the Analyze button. The AI extracts a better title, a short description, detected text, a color palette, and a category.

One of the most useful parts is text extraction (OCR). It works on screenshots, photos of whiteboards, and even handwritten notes. The AI reads the text and gives it back to you as copyable, searchable content.
All extracted text is indexed by full-text search, so you can find images by the words that appear inside them.

The extracted text, description, and color palette appear in the analysis tab, so the analysis remains available when you need without getting in your way.
The detected color palette is shown as visual swatches, and you can click any color to copy its hex value.

All of this data also flows into your image's properties automatically. The Category and Colors properties are populated by the AI, so you can filter and sort your images by what's actually in them.
You can then build queries like "all handwritten notes" or "all blue images" and embed them into moodboards or project pages.

The AI assistant can also use this extracted metadata as context when you chat about your media, without needing to re-read the original image file. It might find connections you did not realize you had. 💫

Image analysis roadmap 🛣️

Media analysis is just getting started.
Next, we’re planning optional automatic analysis for new images (with a setting to turn it off), a rollout to all Pro users, and support for more media types beyond images. Stay tuned!

Aliases for Tags

You can now add aliases to your tags.
This makes it easier to find tags when you use different terms for the same concept, or if you are multi-lingual and you want to refer to the same object in different languages.

Object info sidebar deprecated

The object info sidebar has been deprecated. Its functionality has been redistributed:

• Text stats and created-at info have moved into the object menu:
• Document outline is now a single-purpose floating element:
• Objects inside can now be found in a dedicated section in the new sidepanel:

Default state for backlink sections

You can now choose whether backlink sections start open or closed on object pages. This applies to Backlinks and Unlinked Mentions (Pro) on the web or desktop apps. You can still override the default per object.
There's also a new Backlink content view setting that lets you choose the default view for backlink content. Embed is the default when no option is selected.

Important: Storing Issues in v1.60.1

There is a critical bug in v1.60.1 that in rare cases prevents some objects from being stored to your device. Please update to v1.60.20 on all devices as soon as the version is available.

Community spotlight: Capacities Quick Note Pro for Android

Shout-out to developer Daniel Peck who built Capacities Quick Note Pro, a handy Android app that lets you jot something down on your phone without opening the full app.
Key features:

• Instant capture: type a note and append it directly to your daily note.
• Share from any app: send text from other apps straight to Capacities.
• Homescreen widget: capture without even opening the app.
• Multiple spaces: switch between your Capacities spaces.

The app connects through the official Capacities API, so you'll need a Capacities Pro account to use it. It's available on the Google Play Store for $1.99.

✨ Improvements

• Better triggers for command palette and linking search.
• Better ranking behavior in command palette and linking search.
• Hide Backlinks in main view if they are open in the side panel.
• Unlinked mentions of an object are not limited to 100 elements anymore.
• Better display of the header if you have a single tab only.

🐛 Fixes

• Fixed an issue where starting an AI chat in the AI panel then opening the full chat window would clear the prompt.
• Corrected a misspelled word in the prompt asking users to verify their email.
• Fixed alias linking so it no longer disables page embedding.
• Fixed how tags are shown in embedded task views.
• Fixed the "Learn More" link in the Sharing section which was directing users to an inaccessible page.
• Fixed a bug where text disappears if a bullet is indented right below an image.
• Fixed object type colors in create and link menu not reflecting settings.
• Fixed sending a task to Things not adding a link back to the Capacities content.

Original source

This release brings a redesigned sidepanel, new ways to organize your media, and several quality-of-life improvements across the app. 💫
Check out the video release notes!

A redesigned sidepanel for focused work

If you regularly work with sources, references, or notes side by side, this update is for you. The sidepanel has been redesigned to make it faster to pull up related content while you work.

We've added a static sidebar on the right with quick access to search, AI chat, and backlinks. This is still evolving, and we'd love to hear your thoughts.

Open any object in the sidepanel to see its full content without leaving your current page. This works across all object types. You can even open your object dashboards in the side panel.

You can now browse backlinks directly in the sidepanel too, either within the object itself or in a dedicated backlinks panel. This makes it easier to review what references an object without having to continually scroll up and down your notes.

You can also open multiple objects in the sidepanel at once. Drag any side panel tab into the dedicated area at the bottom to view and resize as needed. Up to three tabs can be stacked in the right side panel.

Hover over a sidepanel tab to reveal the new swap button. Click it to move that content into the main panel so you can continue working on it there.

We also added a shortcut to open and close the side panel. Use Cmd + Shift + Right Arrow on Mac or Ctrl + Shift + Right Arrow on Windows.

The bulk actions menu now lets you select multiple objects and click Open to load each one in its own sidepanel tab. This is useful when you want to review several items at once. You can also click AI Chat to start a conversation with all selected objects as context.

More properties to basic media types

Audio, image, and weblink objects now include a Category property so you can sort and filter them more easily. Weblinks also get a Topic property, and Images now include Colors. These are available to all users.

These properties form the basis of some very exciting updates. 👀 Believers will get an email with more information in the coming weeks. 🚀

Object Sidebar changes

The object sidebar now gives you a quick way to see what's inside a long document. It shows the document outline, referenced objects, and stats in a static sidebar, accessed by the new info-icon.

Click any heading indicator in the document to jump to that section directly. This replaces the previous floating sidebar.

The three-dot menu has moved to the top right of each object rather than the top right of the window. When you have the sidepanel open, this makes it clear which object you're acting on.

New space creation flow & restartable onboarding

If you want a clean start or want to set up a space for a specific use case, you can now choose from a set of templates when creating a new space.

You can also restart the onboarding from Account Settings. This creates a new space and leaves your existing content untouched, so there's no risk of losing anything.

Trash 2.0

You now have full control over what stays and what goes.

• Permanently Delete: Remove individual objects from the trash at any time. By default, trashed content can be restored for 30 days before it's permanently deleted.
• Empty Trash: Delete everything in the trash at once.
• Permanently Delete for deleted spaces: In Account Settings, you can permanently delete all content of a deleted space whenever you're ready, rather than waiting for the 30-day window.

Better search results for the AI assistant

When you ask the AI assistant a question, it now searches your knowledge base using both full-text and semantic search. This means it finds objects that match your words, but also objects that are conceptually related. A search for "dog" will also surface objects about mammals, domestic animals, or pets. 🔍

Results include the object title, type, and a text snippet, so the AI has better context to work with when answering your questions. 🚀

You need to be a Capacities Pro user with AI enabled to use this.

✨ Improvements

• Design of the Object Studio: We made it simpler to create new object types and improved the design to be in line with the new onboarding flows.
• Better wording for states of the export settings based on user feedback.
• Group and sort settings of collections are now also applied in object type dashboards.
• Multiple output properties are now supported in the 'Turn into' dialog. The first matching property wins.
• The 'Turn into' dialog now indicates which properties might be overridden.
• When turning an object into a different type, objects are no longer removed from collections unless the object type actually changed.
• No data loss warning is shown when the target property is set to 'Skip' during a type change.
• Markdown text is now parsed correctly when converting a text property into a blocks property during a type change.
• Text properties can now be mapped to alias properties when changing object types.
• Text properties can now be mapped to icon properties when changing object types.
• Reworked the 'Turn into' dialog layout on smaller and mobile screens.
• Captions are now included in PDF exports.
• Export warnings are now shown at the top of the 'Export' dialog for better visibility.
• Improved how createdAt, lastUpdated, and icon properties are serialized in exports.
• Re-enabled single-dollar math notation ($...$) with heuristics to avoid false matches on currency values like $100.
• Improved tag detection: strings like "Question #2" no longer trigger tag creation.

🐛 Fixes

• Preview mode of Mermaid diagrams now shows the full diagram.
• Critical fix for broken object dashboards.
• Fixed a critical bug introduced with the new sidepanel that broke the UI for some users.
• Indexing bug for backlinks, tags, search, and calendars is fixed. Please wait a few seconds after you download the latest version.
• Fixed: Blocks-type properties no longer incorrectly inherit override state when changing object types.
• Fixed: Exports from tag pages were not working correctly.
• Fixed: PDF exports could display missing objects due to missing user context.
• Fixed: Entity links with nested child content were incorrectly converted to block links.
• Fixed: Formula property values were not being imported correctly.

Original source

With this update, the AI assistant gets richer context, clearer sources, and better grounding in your own notes. 💫
Check out the video release notes!

A Smarter AI Assistant

Our AI assistant can now work with richer context from across your space and show you exactly which notes it used to answer your questions.
Ask it questions about projects, reading, or research and see answers grounded in your own knowledge, across multiple objects.

Richer Context for AI Conversations

We’ve significantly improved how the AI assistant understands and works with your notes.

Bring multiple notes into one conversation
You can now select multiple objects as context for a single AI conversation. Working on a project that spans several notes, documents, and ideas? Select them all at once and ask questions that draw from everything together.
This is perfect for:

• Synthesizing insights across multiple meeting notes
• Comparing ideas from different sources
• Getting a comprehensive view of a topic you’ve explored in several places
  Instead of juggling separate chats or copying content around, you simply pick what matters and start asking questions.

Context currently works for all object types except tags. Support for tags and collections is coming soon. 🔜

Your backlinks become part of the context
When you add objects as context, Capacities automatically adds their linked objects (backlinks) to the context too. This saves you from manually collecting all the related notes first and helps the AI see how things are connected.

Answers Grounded in Your Own Notes

The AI assistant can now read more of your content and clearly show where its answers come from.

Reads deeply into your context
On top of that, the AI can open those linked notes and read their full content. So if your project note links to meeting notes, research, or ideas, the AI can take all of that into account.
This leads to responses that better match your actual writing, decisions, and past thinking.

Finds relevant notes for you
To answer your questions more accurately, the AI can search across your space to pull in relevant notes you might have forgotten about.

Clear sources from your notes
When the AI responds, it now provides clickable links to the notes it referenced.
You can:

• Verify what it says against your original writing
• Discover connections between ideas you hadn’t noticed before
• Jump straight to the most relevant notes for deeper exploration
  This makes the AI more transparent and trustworthy. You can always see which parts of your knowledge base were used to generate an answer.

This Is Just the Beginning

This release is the first version of our more agentic AI assistant.

• We’ll make the AI more agentic so it can better plan its behavior. This means the AI will first research your notes, potentially search the web, and then provide you with an answer.
• We’ll also add more ways for the AI to explore your content, including deeper traversal of your knowledge graph and semantic retrieval across your space.
  We’ll continue to iterate, improve its behavior, and expand what it can do. Feedback from the community is very welcome and will directly shape what we build next.

Our thinking behind AI in Capacities

We’re building AI to support deep, meaningful work over the long term, not short-lived hype. That means focusing on discovery, synthesis, and thoughtful automation grounded in your own knowledge, with clear boundaries so you stay in control.
If you’re curious about how we think about AI in Capacities (beyond features and into long-term meaningful work), we’ve shared more in a new blog post: AI for meaningful work.

Perplexity Reasoning and Web Sources

If you’re using Perplexity as your AI provider, you now get access to Perplexity’s reasoning and web sources directly in your Capacities AI answers.
This combines:

• Your personal knowledge base in Capacities
• Perplexity’s ability to research and reason across the web
  You get answers that can reference both your notes and external sources in a single place, helping you move from question to insight more quickly.
  For example, you could ask: “Summarize my notes on this topic, and include a quick update from the web with sources.”
  For setup instructions, check out our AI documentation.

Data Protection, Transparency, and Control

With these new AI capabilities, we’ve also made changes to how AI can access your data and how you stay in control.

What’s changed

• To enable features like reading backlinks and searching your notes, the AI assistant can now potentially access content across your space. This is what makes the extended context features possible: the AI needs to be able to find and read your notes to help you effectively.
• We’ve also renamed “Only use my keys” to “Prefer my keys”. Some AI features require our infrastructure to work, so we updated the name to more accurately reflect how keys are used.
  If you strictly need to only use your own keys, you can no longer use the Capacities AI features.

Your data remains protected
AI features use enterprise-grade APIs with Data Processing Agreements (DPAs) that ensure:

• No training on your data
• No storage of your content on third-party services
• Processing only – your data is used solely to generate responses

You decide where AI is allowed
You can now activate or deactivate AI for each space individually.
For example, you might:

• Enable AI in your work space to help with projects, research, and writing
• Keep AI disabled in a more private or personal space
  For complete details, see our AI data protection documentation.

Performance Improvements Across the Board

This release also includes a round of under-the-hood improvements across the app to make everything feel faster and more responsive, including:

• Smoother scrolling and rendering in long pages and dense views
• Quicker interactions in menus, pickers, and navigation

New Onboarding

We reworked the onboarding flow for new Capacities users. It’s now more interactive, informative, and fun!

New Docs Page

Our docs have been refreshed with a more modern look and a structure that’s easier to browse, so you can find answers and setup instructions more quickly.

What’s Next?

We recently updated our What’s Next article so you know what we’re focusing on in the coming months.
It’s going to be a huge year and we’re so excited to share with you. 🚀

Improvements and Fixes

✨ Improvements

• ‘Permanently Delete’ for deleted spaces: In Account settings you can now permanently delete all content of a deleted space at any time. By default, a deleted space can be restored within 30 days; after that it is permanently deleted automatically.
• Better ranking behavior in the command palette: Object types are preferred, object type filters have better and stricter matching.
• Removed the border around suggestion items in the date picker on desktop for a cleaner look.
• Adjusted the ‘+’, ‘@’ and ‘/’ menus, so the ‘time’ option does not create a date. Ticket
• We’ve reworked the object menu to structure and group the options more logically.
• The ‘Customize Card View’ section in the object settings has received a visual overhaul.
• Better search matching with a trailing space, e.g. searching for “Object “ in the linking menu returns “Object”.

🐛 Fixes

• Nested blocks could no longer be selected using the mouse. This has now been fixed.
• A rendering issue in menus has been fixed, where incorrect icons were displayed when re-ordering items.
• Fixed numbering issues in the blocks preview to display correctly.
• Fixed: Only first embedded instance appears in backlinks. Ticket
• Fixed links with aliases so they keep the original link text as expected. Ticket
• Fixed rendering problems with drag handles in menu items to improve usability.
• Fixed grouping by backlink objects showing unnamed groups until the backlink is opened. Ticket
• Fixed icon visibility problems during sorting in menus to display icons properly.
• Removed the option to download images that were not uploaded to avoid confusion.
• Resolved a problem with object picker property menus showing inconsistent selections when choosing objects.
• Resolved a problem loading and updating budgets for external AI provider keys.
• Fixed an issue where exporting a tag would result in an empty CSV file.
• Addressed a problem on Windows where file links using file:// paths were not working correctly.
• Fixed a problem that caused issues when selecting blocks inside daily notes.
• Fixed an issue where SVG images for objects were not showing correctly in Chromium-based browsers.
• Resolved a problem where Unsplash images did not appear on page previews in object gallery views and were missing in PDF exports.
• Disable and re-enable block virtualization when using ‘find in page’ to make results more accurate. Ticket
• Fixed localization issues in the offline sync settings to improve language support.
• Fixed issues when selecting multiple blocks across an object.
• Fixed accidental creation of new objects in data views so new items only create when clicking on the intended area. Ticket
• Removed the ‘Collection’ option from the link menu to prevent confusion when creating links.
• Fixed the filter by Object type in queries to work correctly again. Ticket
• Fixed missing content and queries that were pinned to the dashboard so they now appear as expected. Ticket

Original source

🚃 New Storage Infrastructure

We rewrote the whole storage infrastructure within the app. This is a major under-the-hood improvement:

• Substantially faster startup: The app will need to redownload all content with this update, but from then on it should start up much faster.
• Faster content loading: Content loading and calculations are now more efficient.

Please report any issues you experience with the new version on our feedback board. We’d also love to hear any feedback regarding performance.

✨ Improvements

• “Add” option for pinned sections: If you hover over a pinned section you can now click on the “+” button to directly add content to that section
• Better UI for Aliases: Aliases can now be edited and added like tags. By hitting enter you add a new alias. Backspace removes an alias.
• Block linking support for daily notes. Ticket
• Tidy-tree support for Mermaid diagrams. Ticket
• “Capacities needs a restart” now fixed on iOS devices. Ticket
• Better link menu performance. Ticket
• Removed languages for code blocks from create menu for better search results
• The option to create new objects has been renamed from ‘Create Object’ back to ‘New Object’.
• Better behavior of AI auto-filling
• AI auto-filling all properties of an object does not fill aliases anymore as this is usually not intended behavior. You can still use AI auto-fill by clicking on the button for aliases.
• The configuration menu in object dashboards got deprecated as you can now add sections via the “+” button and remove them by clicking on a section when it’s active.
• Customization menu for object dashboard is back
• Better merging of content when working across devices

🐛 Fixes

• We have resolved an issue that resulted in not all objects being shown in data views or all not blocks being visible. Please note that this was only a display issue.
• Unable to view or edit Aliases on iOS. Ticket
• Aliases - show the object type in references. Ticket
• Aliasing text continues after selecting object reference. Ticket
• Hovering over title in order to add description, aliases or autofill etc. does not work. Ticket
• Can’t filter tasks based on the current object. Ticket
• Two issues have been fixed in the ‘Object Selection Modal’, including:
  • The button for selecting files in the ‘Upload’ tab should now work as expected.
  • There should no longer be any rendering issues in the ‘Existing Content’ tab.

Original source

Release Notes

With this update, we’ve completely rebuilt search from the ground up to be faster, smarter, and work entirely on your device. 🔍 🚀
Check out the video release notes!

Introducing Search 2.0

Based on your feedback, we completely reworked the search engine in Capacities. The result is a search that feels instant, works offline, and delivers more accurate results.
The old search relied on server requests and fuzzy matching, which sometimes returned verbose results and required an internet connection. We wanted something better.

A better search engine

• Better matching: We reworked how content is matched and ranked. The search now returns exact matches instead of fuzzy results, making it easier to find what you’re looking for.
• Numbers and symbols: Searching for a price, complex numbers, or symbols now works much better than before.
• Non-logographic languages: The new search engine can handle languages like Chinese, Japanese, Korean, Arabic, and much more. It is not perfect yet but much better than before. Feedback welcome!
• Connected words: Better matching for connected words like “laissez-faire”.

Block-linking search

The search engine now also supports block-linking. Simply type (( and start typing for any blocks content in your space.
This is the reverse of copying a block link and pasting it. You can now simply search for a block.

New search features

We also added new ways to search your content:

• In-page search for tags: You can now, similar to collections or object types, search within tag pages and you’ll get an in-page view of all your search results.
• Exact match search: By using quotes (e.g. “feedback loops”) the search engine will only search for matches of the whole term, not the words within it.
• Anchor search: If you start your search with “^” the search will only return objects, properties, or blocks that start with your search term.
• Better search queries with tags: Tags are now first-class citizens in search queries. You now only see block results that are tagged and match your search criteria.

Performance improvements

We’ve invested heavily in making search feel instant:

• Substantially faster title-based searches for linking, tagging, and the command palette.
• Optimized search performance when using the app for longer sessions.
• Better ranking for titles so the most relevant results appear first.
• Smarter indexing on startup — content will only get indexed if necessary. But you can now also manually reindex your content in the “Offline & Sync” settings if needed.

Breaking changes

• Search queries now do exact matching based on your search term. You can search for a union of search terms by simply adding multiple search terms.
• Command palette does not support filtering for tags anymore.
• Search terms for property name in combination with property value do not work anymore. You can now only search for property values. Although we liked this option, we decided to remove it in favor of more accurate search results.

On top of our new search engine, we’re launching more exciting updates with this release:

Aliases

You can now add an aliases property to all of your custom object types. Aliases are alternative names for your content. Add them by hovering over an object’s title field.
After you add an alias, you can search for it and also reference it when you link to your object.

Mermaid Diagram Support

You can now create beautiful diagrams directly in Capacities! Simply create a code block and select “Mermaid Diagram” as the language.
Write your diagram using Mermaid syntax and see it rendered in real-time.
Mermaid supports flowcharts, sequence diagrams, class diagrams, state diagrams, and much more.

Graph View Tag Network

You can now customize the graph view for tags to only show the tag network. In this view, it excludes all notes and only shows tags based on how they are related to each other.
This is a great way to understand your tagging system and see how your topics connect. You can find this in the graph customization options. Feedback welcome!

Task Inbox Improvements

Task Inbox now only shows tasks with no date and no status: Tasks leave the inbox as soon as you apply a date or status. This allows for workflows like assigning tasks a status like “Some day” or “Deferred” to better work with the “Getting Things Done” methodology.

Capacities API: Breaking Change

We are reworking search in Capacities. This means the API will need to be adjusted as well. There are two changes:

• POST /search: This endpoint will be deprecated on January 30, 2026. Please use the new POST /lookup endpoint instead.
• POST /lookup: This endpoint is new and functions as a title-based search endpoint. The new endpoint has less functionality but is faster and more accurate.
  You can find the updated API documentation here.
  We collected feedback from the community about API use-cases. An extension of the existing API is planned.

Raycast Integration 2.0

An updated version of the Raycast integration went live today. It includes many improvements including:

• Offline support: The Raycast extension now communicates with the Capacities desktop app via a communication token. This means you can use the extension even if you’re offline. Search is still via the server but it’s planned to also be available offline in the future.
• Create Task: You can now create tasks in Capacities from Raycast.
• Search Details: The search results now show the details of the content in a more readable format.
• Better Open Behavior: “Open in App” as default action for opening content in Capacities, “Open in Browser” as fallback action.

Improvements and Fixes

✨ Improvements

• Filtering queries by context: You can select “Backlink objects” in the filter dropdown menus to filter for all objects that have a backlink (or context in case of tasks) to a set of other objects.
• Multiple Pinned Sections 🎉
  • Create multiple custom pinned sections in the left sidebar
  • Customize each with a name and icon
  • Rearrange sections freely via drag-and-drop
  • One-click option to sort items within a section alphabetically
  • Add any object or object type to a custom pinned section using the “Pin to Space” menu option
• Make it possible to rename and delete all pinned sections: You can now rename and delete all pinned sections in the left sidebar.
• Display collections and queries under object types that are pinned to custom sections in the left sidebar.
• Improved Image Viewer (now cross-platform)
  • Smooth zoom between 100% and 500% using mouse scroll, trackpad, or pinch gestures
  • Drag zoomed images (>100%) with mouse, trackpad, or finger
• Embed Resizing Improvements
  • Resizing an embed to full width now correctly syncs across devices
  • Fixed detached resize handlers on image embeds
• Adjusted embed caption color for better visual distinction from regular text blocks
• “Hide dates” option in graph view: Declutter your graph by hiding daily notes.
• Better performance and reliability when filling properties with AI
• Better natural language processing for task titles. It now works better with mentions of “Wednesday”, etc. It also works for dates in the past.
• Option to “Move to today” in today’s calendar view to move overdue tasks to today.
• Tag menu focus behavior improved. Existing tags over other options.
• More items visible in linking dropdowns when no search term is entered
• All linking menus now have loading states so you know if all results are returned
• AI model Gemini Flash 3 added for Google AI Keys

🐛 Fixes

• Unable to search for the content when in the Object Dashboard.
• Tasks not showing on Today Screen.
• Copy and paste tasks deletes the title.
• Fixed an issue where tasks in the Context view wouldn’t show their context group if all tasks shared the same context. Now the context name will always appear for better clarity.
• Corrected the due dates shown for scheduled tasks when sorted by date, resolving timezone related errors.
• Fixed Raycast search fails.
• Fixed Deleting Created Date doesn’t really delete the data, remove this warning.
• Additional mobile stability improvements
• Resolved rerendering and flickering when navigating in the calendar
• Help chatbot window no longer cut off on mobile devices
• Restored autofocus functionality in AI chat
• Fixed various bugs related to pinned sections on mobile
• Potential fix for blank page issues on certain mobile devices
• Various fixes for media embed resizing
• Layout fix for task card view (prevented overflowing context properties)
• Mobile: “Remove” button now works correctly when clearing a task date
• Date picker in task management no longer closes unexpectedly when selecting a month
• Fixed inability to remove the first context added to a task
• Prevented completed (done) tasks from being accidentally moved to other status groups
• Fixed an issue where many objects were duplicated unexpectedly, causing problems with backlinks and data consistency.
• Resolved a critical bug that caused nearly all items in the database to be duplicated after the last release.
• Fixed a major problem in version 1.57.13 that caused most objects to appear in double and lose backlinks when duplicates were deleted.
• Addressed an issue where objects duplicated randomly across several different types.
• Fixed issue with temporary property creation in change type modal.
• Fixed search results to no longer show trashed items, keeping search results clean and relevant.
• Weblink images are not displayed - fixed.
• Web embeds are not working - fixed.
• Tweet embeds from X/Twitter are not working - fixed.

Original source

Introducing Contextualized Task Management

This might come as a surprise but task management is here. ✅ 🚀
We spent countless hours in front of whiteboards, drafts, and ideas over the last months to come up with task management “the Capacities way”. Over the last few weeks, we secretly worked on it behind the scenes and are super excited to launch it for Capacities Pro and Believer today. 🧑‍💻 📚

Everyone needs a system to help them get things done. Having that system embedded directly in the context of your work, without the need for any setup, is a unique approach that other task managers can’t offer. We’re super excited to introduce contextualized task management in Capacities.
The principle is simplicity: create a task from anywhere, find it exactly where you need it, and use built-in views to execute on what matters.
No setup required, it just works. Let us show you how.
📺 Watch this video or read on for more information.

Create tasks from anywhere

Tasks can be added from almost anywhere: inside your notes, through the command palette, the mobile app, the sidebar, or Raycast. A new lightweight modal makes creating them quick.
When creating a task, you can set three key details:

• Where it belongs: the note, project, or object it should show up in
• What its priority is
• When you intend to do it: a day, week, or specific time
  Capacities organizes the task for you from there. You’ll find it where you need it.

Group all tasks in your notes

Whenever you create a task in a note, a dedicated task tab collects them for you.
Several views come out of the box, so you can see your tasks in the way that’s most useful. Again, no setup is required.
This includes a Kanban view that lets you review all open tasks in the context you’re working in and see the next step in the process.

See your tasks in the calendar

Capacities’ central calendar is a core part of the app, and tasks now appear there too, organized by when they’re scheduled.
You can:

• View all tasks for today
• See upcoming tasks for the week or month
• Quickly reschedule tasks that didn’t get done
  This turns the calendar into a simple place for daily planning, without requiring configuration.
  It also means you can see all the tasks done simply by going back in time in the calendar.

Get a full overview when you need it from the task tab

For a broader perspective, the task tab in the left sidebar gives you an overview of all your tasks.
Capacities comes with all the views you need, but of course, you can create your custom task queries on top. Capacities provides:

• Inbox: All tasks without a scheduled date. You can use this to schedule the tasks you created.
• Today: Are all tasks scheduled for today, grouped by status and ordered by priority. You can also see this in the calendar.
• Scheduled: All tasks that have been scheduled, grouped by their date property. Overdue tasks will be at the top for you to reschedule.
• Context: All tasks grouped by the objects they relate to.
• Open: All undone tasks ordered by their priority and status.
• Completed: A log of all tasks you marked as done, ordered by when you completed them.

Philosophy

Tasks in Capacities are simple and actionable. We intentionally designed it so you can do your best work, and Capacities will help you with that. We call it “contextualized task management”. You can read more about why we built task management this way in our blog post.

Roadmap for Task Management

The current version of task management is a “Version 1”. We have ideas and plans to extend task management in Capacities. You can read more about our ideas in our documentation.
In general, if you want to know everything about task management in Capacities, you can read our task management documentation.

On top, we’re launching more exciting updates with this release:

Two themes for Capacities

• Introduced two brand-new themes to give you more control over your workspace:
  • Light Dimmed – A softer take on the classic light theme with reduced contrast and gentler colors. Perfect for long note-taking sessions when you want a light interface without the harsh brightness.
  • Dark – Our new default dark theme! It’s noticeably darker and richer than the previous default, offering deeper blacks and a more modern, immersive feel—especially great in low-light environments.
• Renamed the previous default dark theme to Dark Dimmed for clarity. If you loved the old dark look, it’s still here—just select “Dark Dimmed” in your theme settings.
  These updates make it easier to find a theme that feels just right for your eyes, mood, and lighting conditions. Happy note-taking! 🌙✨

Graph View Customization

Your graph views can now be customized to your preferences! You can click on the slider icon in the graph view to:

• Hide objects of a certain type: Select object types like images, daily notes, or tags to not see them in the graph view.
• Always show the simplified graph: If you prefer to have a cleaner graph view, nodes will only be shown as colored dots
• Hide objects with many links: Large notes can clutter your graph. This option will hide all nodes with more than 25 links.
  We also added various improvements for the graph view:
• Better edge labels in graph
• Added support to show collections in the graph view so tags inside collections now appear.
• Improved graph view to better handle siblings by splitting links going both ways.

Resizable Media Embeds

Working with media just got even easier. You can now easily resize your embedded media objects in your notes.
You will see options for how to align your media, and you can add a caption:
This gives you more control over the layout of your notes.

Highlight Improvements

Based on community feedback, you can now customize highlights from Readwise, Kindle, or the web in the Editor settings. You can choose from:
Default: Highlights are displayed with a subtle background color.
Quote: Highlights have a quote bar on the left side.
Minimal: Only show the highlight menu on the right.
You can build your highlight systems around this.

What’s Next?

2025 has been a huge year for Capacities. We’ve added major new features and continued polishing many aspects of the app. We’re so excited to continue this in 2026.
After the holidays, we will share another “What’s Next” article, building on our most recent one. There are still some projects on there that we’re working on and that will launch next year.

From Us

Here’s our latest content:
📺 Overview of AI in Capacities
📺 Web Highlighting in Capacities
📺 What to Do After Highlighting: A Reading Workflow

Improvements and Fixes

✨ Improvements

• New AI Provider: Perplexity is now available as a custom AI provider.
• New Mobile Calendar View: We improved the daily calendar view on mobile to make it more accessible and easier to use on mobile.
• Better edge labels in graph: If there are labels in both directions on an edge, you now only see one.
• More sort options for date properties: You can now sort date properties by “Exists”/“Not exists” to show objects at the top that have dates set or not set.
• Group sort option for “Backlink objects”: You can now group all objects by the objects they are backlinked to by the date these objects were created.
• Added keyboard shortcut Cmd/Ctrl + T to open a new tab in the desktop app for quicker navigation.
• Added support to show collections in the graph view so tags inside collections now appear.
• Improved graph view to better handle siblings by splitting links going both ways.
• Updated the graph views to hide large nodes and show minimal nodes for a cleaner display.

🐛 Fixes

• Fixed an issue where the desktop app showed “View is not ready yet” even though content was synced and available.
• Resolved problem where deleting an object caused the opened editor to close unexpectedly.
• Corrected shortcut keys on macOS so commonly used shortcuts like Command+U and Control+Command+H work properly.
• Fixed an issue preventing users from deleting default object types.

Original source

Our Readwise Integration is here!

Readwise Integration Now Available!

You can now bring your Readwise highlights into Capacities to turn them into connected knowledge. 🚀

How it works: Instead of automatically importing all your highlights, you choose which content to bring into Capacities with their highlights. This helps reduce information overload and help you find the useful highlights quicker!

Why we built it this way

We know this looks different from other integrations you might have used. We want Capacities to help you think, process, and connect, not to contribute to information overload.

Importing everything creates a backlog that feels overwhelming. By being intentional about what you bring in, you can build a sustainable reading and note-taking workflow.

The goal isn’t to capture everything. It’s to find real insights and create connected knowledge.

📺 Watch the setup video

📖 Read more about our approach

Reading Worklow

If you’re wondering what such a workflow would look like, then check out our latest tutorial or blog post!

Faster text editing

We improved the performance of text editing, especially when working in large documents. You should notice a significant improvement when opening notes and while editing text. Performance improvements are top priority for us and we will continue to work on them in the upcoming releases.

v1.55.7 (2025-11-07)

✨ Improvements

• Added anchor elements to keyboard shortcuts to improve expanding block selection with Shift and arrow keys.

🐛 Fixes

• Fixed the autofocus behavior of the main button in our modals to restore expected focus functionality.
• Added proper bottom padding to data views to ensure content is fully accessible without clipping.
• Corrected cursor placement issues when deleting formatted text, improving text editing reliability.
• Resolved an app freeze that occurred after converting the type of an object; typing works correctly without needing a refresh afterward.
• Fixed input issues within notes containing embedded weblinks to allow proper text entry.

v1.55.6 (2025-11-04)

✨ Improvements

• New property creation modal is now more forgiving: the modal no longer closes if the cursor strays slightly outside while typing or configuring a new property. This reduces accidental cancellations and improves keyboard-first workflows.
• Hide object types from calendar sidebar: In the object settings you can now select to hide an object type from the daily calendar sidebar on the right. Head into your object type settings > Calendar Settings > Hide object in calendar sidebar to set this up!
• “Create new Tab” button in space tabs: If you have multiple tabs, you can now simply add a new one by click on the plus button.
• The date picker now automatically opens for the ‘date’ and ‘time’ options in the create menu (‘+’', ‘/’) for faster input.
• Added quick commands to the ‘/’ and ‘+’ menu: ‘tag query’ and ‘search query’ for faster inline query creation.
• When indenting under a heading, new blocks now default to normal text instead of inheriting the heading level.
• Updated the colors of the find-in-page popup on Desktop to match the new dark mode.
• Updated to the latest emoji set baseline.
• Added anchor elements to keyboard shortcuts to improve expanding block selection with Shift and arrow keys.
• Less horizontal paddings for Kanban columns
• Better performance in table view and data views in general.

🐛 Fixes

• Fixed the chatbot input field being hidden while typing on mobile devices.
• Resolved an issue where adding PDFs to a new collection stopped working.
• Fixed labels in dropdowns on iOS where the bottom border was cut off or invisible, improving visual style.
• Corrected inserted PDF links that were converted into PDF objects without content, ensuring proper PDF display.
• Fixed an issue where table view layout settings would reset after reloading the app in queries, improving the persistence of your view configurations.
• Resolved a bug that prevented reordering of columns in the Kanban view, making columns draggable again for better customization.
• Removed an unnecessary separator line in the mobile date picker for a cleaner appearance.
• Prevented the ‘Properties’, ‘Content’, and ‘Graph’ options from displaying over the scrollbar on mobile devices.
• Corrected the direction of arrow icons in collapsible sections like backlinks to improve visual clarity.
• Fixed layout issues with search bars by ensuring the clear button only appears when there is text, reducing unwanted white space.
• Restricted the Cmd+A shortcut in data views to avoid unintentional object selection without proper constraints.
• Changed ‘Copy as Markdown’ shortcut to ‘Cmd+Shift+C’ on macOS and ‘Ctrl+Shift+C’ on Windows to avoid browser conflicts.
• Improved nested block selection reliability in the editor. Selecting parent and child blocks now behaves predictably.
• Resolved an issue where creating inline links while typing could replace the first character unexpectedly.
• Fixed ‘Search Query’ not opening an input field and Capacities being unresponsive afterwards.
• Resolved card preview overflow in gallery/card views when a block reference was included.
• Corrected the close ‘X’ hitbox in find-in-page popup on Desktop, so interactions map accurately.
• Restored automatic theme switching with the system theme without requiring a reload on Desktop.
• Weblinks saved via the Web Clipper now show proper preview images instead of only a blurred background.
• Mobile: Fixed cover images not being editable on mobile.
• Fixed the autofocus behavior of the main button in our modals to restore expected focus functionality.
• Added proper bottom padding to data views to ensure content is fully accessible without clipping.
• Corrected cursor placement issues when deleting formatted text, improving text editing reliability.
• Resolved an app freeze that occurred after converting the type of an object; typing works correctly without needing a refresh afterward.
• Fixed input issues within notes containing embedded weblinks to allow proper text entry.
• Share and export menus working again.
• Fixed a bug where line wrapping behavior in code blocks could not be changed on mobile.
• Fixed a bug where toggles and to-dos could not be tapped on Android
• Fixed an editor bug that occurred when working with the sidepanel or preview model (cursor would not update correctly and link dropdown menu failed),
• We have resolved an issue that caused major performance issues on MacOS 26 Tahoe.
• Proper coloring of group headers (when group-by is active) in data views. For example, when a data view is embedded inside a block with colored background, the same background color is used.
• Hovering over math blocks now reliably shows the block handle, making it easier to select and move these blocks.
• Corrected rendering for ‘indigo’ and ‘teal’ label color options so they no longer appear gray.
• Fixed a large gap appearing between the menu bar and the on-screen keyboard on tablets. This restores proper editor positioning and tap targets.⁠
• It is now also possible to select text and position the cursor on touch devices using a mouse.
• Corrected inline alignment for text, mentions, and LaTeX so mixed inline content lines up consistently across platforms.
• Resolved broken PDF acquisition from web URLs. Creating embeds from pasted PDF links now displays the document correctly across objects and devices.
• Kanban rows snap back to their old position after moving.
• Displaying shared collection works again
• Better Tag Presentation
• Toggle wide cover image in a property - not always available
• Fixed the Escape key in the preview modal so it now closes the view instead of selecting the current block when block selection is active.
• Resolved an issue where child blocks were deleted unexpectedly when the parent block was merged with the preceding block. Now, child blocks are preserved and correctly placed below the current block.
• On Android: when focusing the daily note programmatically, the keyboard was not opened.
• Fix: selecting a large number of blocks by drawing a rectangle stops working after a certain number of blocks
• Fix: scroll position jumps when editing highlight from Readwise

Stay updated

Subscribe to our newsletter so you won't miss out. We will send you updates via email about Capacities and the space of personal knowledge management.

We will use your email address for sending you our newsletter. You can change your mind at any time by clicking the unsubscribe link in the footer of any email you receive from us. For more information see our Privacy Statement.

Original source

Choose your own AI Model Provider

You can now use Capacities AI with other model providers than OpenAI. You can now plug in your own API keys from OpenAI, Anthropic, Google, Mistral, or xAI.

We changed our AI infrastructure to a more versatile and flexible system:

• You choose where you send your data: By selecting a model provider and making it the default, you are in full control over where your data is sent.
• Benefit from the advantages of different model providers: Different providers have different strengths and weaknesses. By being able to choose, you can benefit from the best of each provider.
• Get access to the latest models: With the new infrastructure we can quickly add new models as soon as they are released by the provider so you can use them in Capacities.

For the setup, head to the AI settings in Capacities then simply choose the models you prefer, set your own spending limits, and even assign keys to specific spaces. Just add a key in Settings > AI Assistant and you’re ready to go. 🚀

⚙️ For more information on setup, check out our docs or this video.

Kindle Integration

We’re excited to launch our Kindle import integration!

You can now bring your Kindle highlights directly into Capacities via our email integration by emailing your highlight exports to [email protected].

The Kindle format preserves your book’s chapter structure, which can be super useful when reviewing your highlights, particularly in long books.

📺 For a 30-second intro, check out this video
⚙️ For the setup, check out our docs
💡 To understand the difference between our recently launched Readwise integration and Kindle imports, check out the docs

Web Highlights 2.0

Web Highlighting got a big update in Capacities!

Thanks to Marius, we now have a Web Highlights 2.0 integration that combines the power of Web Highlights with the power of our highlight blocks.

If you highlight a lot of web pages, check out the latest iteration of the integration.

📺 To see this in action, check out this video
⚙️ For the setup, check out our docs
💡 To make a choice between the Capacities Web Extension, our Readwise Integration, and Web Highlights, check out our web extension documentation

Raycast for Windows

Raycast is a very powerful app based around a launcher that can supercharge the apps you interact with. Thanks to Syed, it’s now available on Windows too!

With it, you can search your notes, add to your daily note, and save web links without having to open Capacities.

Head over to the Raycast store to download it. 🚀

From the team

We’ve published lots of content recently!

• 📝 Using Collaborative Tools Alongside Capacities: Our Two-Phase Workflow
• 🤔 What challenges do we face in growing Capacities?
• 📺 How Réka Uses Sublime & Capacities
• ❓ When To Create A New Object Type
• 📺 How to Export Your Highlights to Capacities from the Readwise team!

Improvements and Fixes

On top of our new integrations, we’ve also made some improvements and fixes to the app. We hope you like them!

Improvements

• We have adjusted the criteria for the media loading state under ‘Settings > Offline & Sync’ so that the severity is not as high if only a handful of media files fail to load.
• Moved the “delete object type” option from the three-dot menu to the settings for clearer organization.

Fixes

• Fixed issues with text inputs in editor menus to prevent conflicts after using keyboard shortcuts.
• A problem has been fixed where some long-running tasks (e.g. a full export) were automatically canceled after a certain amount of time. These timeouts have now been adjusted.
• Fixed several accessibility and keyboard shortcut issues in the editor panel for smoother navigation.
• Resolved an issue preventing import of .md files on older macOS versions by updating accepted file extensions in the import menu.
• Corrected the create menu so that block creation options such as headings always appear at the top and sorting does not push them down.
• Fixed a broken ‘Learn More’ link in the documentation to ensure it directs correctly.

Original source