Tana Current October 2026

A larger team for Tana Outliner, improvements to everyday work, and a look at bringing your own agents into meetings with the new Tana.

Tana Current, October 2026 Team Update artwork with both product logos and three speaker portraits on a light background

Two products, more ways to work

October’s Tana Current covered both products: a live demonstration of agents working inside a meeting, a larger team for Tana Outliner, and a walkthrough of recent releases and work in progress.

For Outliner users, there’s plenty to explore. Grouping, hidden fields, dashboard layouts, and time zones have all received attention. Grim, who built the original knowledge graph prototype, is returning to lead Outliner.

Olav and Brage also shared the direction for the new Tana: bringing the agents you already work with into meetings, so they can act on the conversation while it’s happening.

This recap covers the October 20 session. Where features have shipped since the demonstration, we’ve noted that separately below.

Tana Outliner: a change in product leadership

Grim built the first Outliner prototype, including the data structure and primitives that make it possible. In the session, we announced his return to lead Outliner, building on the work Mark and Odin have been driving. With more engineers joining, we’re increasing the team’s capacity to improve everyday work and develop the knowledge graph further.

The priorities are practical: smoother everyday use, support for larger workspaces, and better ways for people and agents to work with structured information.

Recent Outliner improvements

You can find release notes inside Tana Outliner in the What’s new tab under notifications. Mark walked through several improvements during the session.

More control over grouped views

Groups can be collapsed, show item counts, and hide automatically when empty. You can also exclude individual groups while keeping Auto-include new values enabled, so new options still appear as your system grows.

These changes help you keep a view focused without continually adjusting its configuration. See the October 12 release notes.

Hidden fields that take up less space

The Collapse hidden fields preference reduces hidden-field buttons to a single line you can expand when needed. Hold Option on Mac, or Alt on Windows and Linux, and click a supertag field’s bullet to hide it. For persistent behaviour, use the field’s hide conditions.

Lanes for dashboards

Lanes arrange a node’s children side by side, with other views nested inside. You could put projects, actions, and areas next to each other, then expand a search without pushing the next section down the page.

Choose View as → Lanes to try it. Lanes and the hidden-field controls are covered in the October 9 release notes.

Time zones in date fields

Dates can carry a time zone, with a preview of the converted time. You can set a display time zone in your preferences.

The walkthrough used a flight from Adelaide to Oslo to show how dates work when the start and end are in different zones. Checking a time in another zone is useful when planning with people elsewhere.

Instructions for agents

Mark also highlighted how AI instructions on supertags and fields can guide agents working with the structure you’ve built. This is part of the work on Outliner’s MCP, alongside improvements to contextual content and performance in larger workspaces.

From the demonstration to release

Odin showed changes at different stages of development. The export work was rolling out during the session, while reminders were demonstrated in a staging environment. Code documents and readable workspace export are now included in the November 2 release notes.

A simpler JSON export

Tana Intermediate Format 2.1 gives you readable content and schema that you can import into a new workspace. Choose Export workspace as JSON, then select Tana Intermediate Format 2.1 once the workspace has finished loading.

The format carries structured content, including fields, supertags, and references. Importing creates an editable copy with fresh node IDs; references outside the exported workspace still point to their original targets.

It is a lossy format. Views, automations, workspace settings, membership, and appearance metadata are not restored, and attachment files are linked rather than copied. The separate Tana JSON option retains the full stored representation for technical use; it cannot be imported through the TIF importer.

Reminders on date fields: a preview

The staging demonstration showed a Remind me setting in a supertag’s field configuration. A reminder on a task’s due date appeared in notifications, and an Android push notification opened the relevant node.

Browser and Android notifications were discussed as the first supported surfaces. The Android update had not been released at the time of the session, and iOS support was planned to follow. These were preview and roadmap updates, rather than an announcement of general availability across those platforms.

Better pasting and code documents

Paste options let you change how pasted content is interpreted after inserting it.

Odin also previewed code documents. Since the session, these have shipped: code you paste, import, or add with Create code document opens with a code editor, syntax highlighting, and a language picker.

The demonstration showed titles, code longer than 100 lines, and Tab indentation while editing. Odin explained that code documents are stored as attachments to keep long pieces of code out of the workspace’s main content.

What we’re working toward

The session also covered deeper architecture work: server efficiency, loading performance, support for larger workspaces, and reducing dependence on an internet connection.

The priorities discussed were:

  • Faster loading on desktop and mobile
  • More efficient search and server operations
  • Better support for larger workspaces and longer offline use
  • Reminder support across browsers and mobile apps
  • Making Outliner’s MCP easier to discover and connect to

These were priorities discussed in October; no release dates were announced in the session. Longer offline use was an architectural direction, rather than an announcement of a new offline limit.

A ChatGPT store listing was also discussed as a future route to Outliner’s MCP. That is separate from the documented custom MCP connection, which is available now. The hosted server is in beta and requires a paid plan; the local server remains free. We are not announcing a store listing here.

MCP is changing how people use Outliner

Matt highlighted how much MCP has changed the way people use Tana Outliner over the past six months, particularly in the last three. Agents are becoming a regular way to create content, retrieve information, and work with the knowledge people have built in their graphs.

In an internal snapshot Matt cited during the session, 94% of newly created nodes came through MCP. He shared it as an illustration of how usage is changing. It describes node creation in that snapshot, rather than the proportion of people using MCP.

Asking questions across a graph and pulling out relevant information through an agent is already part of Matt’s everyday work. It’s something people can explore through MCP, rather than a single built-in “talk to your whole graph” feature in Outliner’s AI chat.

Mark has been exploring these workflows in Systems Lab sessions. We’ll continue sharing examples to help people discover what they can do with MCP in their own workspaces.

The new Tana: bring your own agents into the meeting

Much of a meeting’s context lives on screen. A design someone points to, a slide they’re revising, or an issue they demonstrate can be difficult to reconstruct from a transcript alone.

The new Tana gives agents context from what’s said and shared on screen. It is a separate product from Tana Outliner. You can bring agents such as Claude and ChatGPT into the meeting with the connectors, skills, and memory you already use with them.

From a conversation to a slide

In the demonstration, Brage shared Outliner material on screen, then asked Claude to use it to add a slide to a presentation in Paper. Claude worked in the design tool while the meeting continued, using the visual context Tana had captured.

The result still needed refinement. Brage could point to what needed changing and continue the conversation. The useful part was directing the work from the meeting, with the agent following what was discussed and shown.

Your own agent, your own chat

Each person can work privately with their agent during the meeting and choose when to share the result. You can ask for help or act on a discussion without interrupting everyone else.

The session also showed voice interaction and Tana’s own agent, alongside the option to connect an external agent.

Working through something together

This applies to solo sessions, too. Olav described going through a deck with an agent, scrolling through the slides and discussing the opening, the argument, and what needed work. The agent could follow the material as the conversation developed.

Meeting context can also help people catch up across time zones, understand what happened, and see where they need to contribute.

What’s ahead

Both products are important to us. The new team structure gives Outliner more capacity, while the new Tana brings agents closer to the conversations where work happens.

During the session, we discussed recruiting another group of test pilots for upcoming Outliner work and holding a founder AMA before the end of the year. Check the community for current announcements.

Keep sharing feedback and ideas. They help us understand where the next improvements will matter most.

Watch the October Tana Current recording on YouTube

Next Tana Current: check the community for the latest announcement.

Explore further

Tana Current October 2026 - Tana Outliner