🔍 Read the full analysis: One Desk, Multiple AI Agents: My Thrymvault Setup on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Thorsten Meyer says his news network is building Thrymvault as a shared workspace where an editor can track AI agents’ newsroom work, steer stories and trace changes. It uses approved agent applications and does not allow agents to delete content, but parts of the setup—including a standby machine and hourly backups—are still planned.
Thorsten Meyer has described Thrymvault, a workspace he is building for the AI agents used across his StrongMocha News Group, as a shared newsroom desk for tracking stories, instructions and changes. The November 13 account sets out how the system is designed to work, while distinguishing features available now from those still being rolled out.
Meyer says the newsroom previously handled work through a chat window and a task tracker designed for code work. On one day this month, he writes, the team used that arrangement to steer nine articles: agents gathered sources, drafted copy, made graphics and prepared drafts. But story status was not visible in one place, revisions depended on messages reaching the right agent, and finished explainer pages could not be shared as links.
Thrymvault is a workspace for pages, databases and boards, built and hosted on the group’s own machines. Meyer compares its general shape to Notion, while emphasizing that agents access it through limited interfaces. The available routes he describes are an MCP server with 20 tools, a REST API and a command-line tool. The agent guide is published as plain text and as a machine-readable manifest, generated from the same tool catalogue as the server.
In the proposed newsroom workflow, a topic appears as a board card. Meyer decides whether it advances; moving it to a “Kit” stage prompts an editorial agent to gather primary sources and prepare a source kit. An agent then drafts the story and records its status, draft link and open questions on the board. Editors can comment on the card and mention the relevant agent application to request changes. The agent’s reply and edits are associated with that application, and the card can ultimately hold the published link.
The shared desk · animated
How a story moves through the shared desk
One card per story. The editor moves it and asks questions in its thread; the agent does the work, keeps its status current and answers there. Every agent change is signed with its application.
Editorial board
Thread on the card
Editor@Editorial agent shorter title, no colon
AgentDone: title shortened, draft updated.
via Editorial agent
Signed history
- Moved to Kit · by the editor
- Status and draft link updated · via Editorial agent
- Comment mentions Editorial agent · by the editor
- Reply in the same thread · via Editorial agent
- Approved and scheduled · by the editor
- Published link added · via Editorial agent
Agents add and edit. They cannot delete anything.
Source: Thrymvault agent integration guide and rollout plan, November 2026. Shows the desk as designed for StrongMocha News Group; the board templates and the pick-up of mentions are still being rolled out.
Animation: Thorsten Meyer AI
A Traceable Desk for Agent Work
The system addresses a coordination problem that grows when several agents work on the same stories: a chat log can contain instructions, but it does not necessarily provide a durable view of each story’s status, decisions or revision history. A shared board could give editors a clearer place to review work and direct changes, while keeping notes attached to the relevant item.
Thrymvault’s access controls are central to Meyer’s argument. Agents enter only through applications approved by a workspace owner; each change identifies the application that made it, and agents cannot delete content. Meyer says an application key acts with the rights of the person who created it, within one workspace, and cannot exceed the rights permitted for that application. For readers, the intended benefit is a more traceable path from draft to publication that could help a newsroom investigate and correct mistakes. The article describes that rationale, not evidence that the system has already improved accuracy or reduced errors.
The approach also shifts editorial instructions out of transient conversations. Meyer says decisions such as a new headline rule would be recorded in a “Decisions” database with the wording, date, scope and a link to the affected rule. That could help agents and editors find the basis for a change later, though the description does not establish how consistently those records are maintained.
Products Worth Considering
Logitech MX Master 4 Ergonomic Wireless Mouse with Haptic Feedback, Graphite
Apple MacBook Neo 13-inch Laptop with A18 Pro Chip, 8GB Memory, 256GB SSD, Liquid Retina Display, and 1080p Camera — Blush
FLEXISPOT Electric Adjustable Workbench, 48 x 30 in, Maple
As an affiliate, we earn on qualifying purchases.
From Chat Threads to Story Cards
Meyer’s account contrasts chat-based coordination with a newsroom desk organized around stories, statuses and attached notes. In the earlier setup, he says, he had to ask an agent where a story stood, and a headline revision depended on the right agent noticing a chat message. The problem was not that agents could not produce material, but that their parallel work lacked a shared editorial view.
Thrymvault is presented as a workspace built for both people and software agents rather than a general-purpose chat. The described workflow uses card movement to signal editorial progress, but Meyer clarifies that moving a card is an agreement for agents to follow, not a software switch that automatically starts a task. If a card is marked “stopped,” agents are meant to leave it alone. The distinction matters: the process depends in part on agents following instructions, not solely on enforced automation.
The article also identifies infrastructure that is not yet in place. Thrymvault runs on the group’s own hardware, while a second machine on standby and hourly backups are described as the next step rather than live capabilities. The account is therefore both a design description and a status report on an evolving internal system.
“That is the point where an AI-run newsroom outgrows chat.”
— Thorsten Meyer, describing the newsroom’s earlier workflow
Rollout and Safeguards Still Developing
Meyer’s article does not provide an independent evaluation of Thrymvault or evidence that every step in the described workflow is already operating in routine publishing. It identifies some current capabilities, including the agent interfaces and the restriction on agent deletion, but does not give a complete feature-by-feature rollout status. The standby machine and hourly backups are explicitly not live.
Other practical details remain unspecified, including how the system handles an agent that ignores a card status, how access is revoked after a key or application is no longer approved, and what recovery procedures apply if the self-hosted workspace becomes unavailable. The article also does not report security testing, uptime figures, newsroom-wide adoption or measured effects on publishing speed and error rates. Its claims about traceability and easier correction describe intended benefits, not demonstrated outcomes.
Standby Hardware and Rollout
Meyer identifies a second machine on standby and hourly backups as the next infrastructure step. He does not give a deployment date. The newsroom’s described workflow would also need to be rolled out across story stages, agent applications and editorial decision records; the article does not specify a timetable or say which parts are already in regular use.
The next useful progress update would clarify what has moved from design to daily operation, whether the backup and standby setup is active, and how the team tests permissions and recovery. Until then, Thrymvault is best understood as a developing newsroom workspace whose stated aim is to put agent work, editorial direction and change records in one place.
Key Questions
What is Thrymvault?
Thrymvault is a self-hosted workspace for pages, databases and boards that Thorsten Meyer says he is building for his newsroom and its AI agents.
How do AI agents access the workspace?
Meyer describes three routes: an MCP server with 20 tools, a REST API and a command-line tool. Agents are allowed in through applications approved by the workspace owner.
Can agents delete newsroom content?
According to Meyer’s description of the agent integration, agents cannot delete anything. Changes are attributed to the approved application that made them.
Is the backup system already running?
No. Meyer says the workspace runs on the group’s own machines, but a second machine on standby and hourly backups are planned next steps, not live features at the time of publication.
Does Thrymvault automatically move a story through production?
Not necessarily. Meyer says moving a card between stages is an agreement agents are expected to follow, not a software switch that automatically starts work.
Source: ThorstenMeyerAI.com
