This Tana Outliner Systems Lab shows how Remote MCP connects your Outliner graph to the new Tana. Outliner supplies the structured context; new Tana supplies the meetings, skills, and agents that use it. Approved meeting outcomes can then return to Outliner as proposed updates or new records.
For the one-time connection steps, start with Connect Tana to Tana Outliner. This session focuses on what to build once the connection works.
Intro [0:00]
The session builds on the first Systems Lab and moves from connecting the two products to using them as one working system. Tana Outliner remains the structured history and destination for resulting work. The new Tana is where its linked meetings and agents run.
Remote MCP Recap [2:30]
Remote MCP gives the new Tana permissioned tools for searching, reading, and writing in your Outliner graph. Because the server is hosted, the new Tana can reach it without the Outliner desktop app running. The connection appears there as a linked skill whose individual tools you can allow or deny.
Demo Scenario Overview [8:35]
The demo follows one new-Tana meeting across its full lifecycle: prepare from existing Outliner context, retrieve more context while the conversation is happening, and propose updates after the call. The point is not a one-off prompt. Each working prompt becomes a new-Tana skill that can be run consistently or handed to an agent.
Skill: Meeting Prep From Outliner [10:40]
The first new-Tana skill gathers relevant context from Outliner and turns it into meeting prep. It can search beyond a single page, follow the references around the subject, and assemble the open threads into a brief you can use in the meeting. Saving the workflow as a skill gives every run the same instructions and output shape.
This follows the same progression as Walk into every meeting prepared: get one result right in new-Tana chat, save the method as a skill, then automate it with an agent.
Agent: Daily Meeting Prep From Outliner [18:30]
The meeting-prep skill becomes an agent that runs across the day's calendar. Instead of naming and preparing meetings one at a time, the agent applies the same process to each relevant meeting and leaves the preparation ready for review. Add a schedule to make that preparation happen automatically at the start of the day.
Skill: Grab Relevant Outliner Context During Meeting [27:30]
Remote MCP also works inside the new Tana meeting's Private Meeting Agent. A retrieval skill can search Outliner when the discussion surfaces an older decision, project, or note, then bring the relevant context into the current meeting without switching tools. Naming Tana Outliner in the request, or invoking its skill with /, makes the intended source explicit.
Skill: Create and Update to Outliner After Meeting [34:25]
After the meeting, another skill reads the outcomes and proposes what should change in Outliner. Existing records can be updated and genuinely new records can be created in the hierarchy and structure you specify. Name the destination, tag, and fields in the skill so the AI does not have to guess where the work belongs.
Keep this as a proposal-first workflow. Review the suggested changes before applying them, especially when several Outliner records may match the meeting context.
Agent: End of Day Meeting Review [40:30]
The final agent applies the post-meeting skill across the day's meetings. Its job is to find the outcomes that should update Outliner, prepare the changes, and bring them together for review. Scheduled alongside the morning prep agent, it closes the loop: Outliner context enters each meeting, and approved meeting outcomes return to Outliner at the end of the day.
Discussion [41:50]
The session closes with discussion of the workflows and how to adapt them. The durable pattern is the same throughout: connect a source with MCP, make one grounded workflow reliable as a skill, and give that skill to an agent only after its inputs, destination, and review step are clear.