🔍 Read the full analysis: How To Set Up A Shared Desk For AI Agents on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Thorsten Meyer says his newsroom is building Thrymvault as a shared workspace where editors can track AI-agent assignments, drafts and decisions instead of relying on chat. The system is self-hosted and uses approved applications to identify changes, but some planned infrastructure, including standby hardware and hourly backups, is not live yet.
Thorsten Meyer has outlined a shared newsroom workspace called Thrymvault for coordinating AI agents, replacing chat-only task handoffs with boards, attached reports and traceable changes. The system is being built and hosted on the publisher’s own machines; Meyer says a standby machine and hourly backups are planned but not yet live.
Meyer said that during one day this month, his team used a chat window and a task tracker designed for coding to guide agents through nine articles. The agents gathered sources, drafted copy and prepared graphics, but he had to ask for status updates, send headline instructions through chat and lacked a shareable link to one completed explainer. He presents those problems as the reason the newsroom needs a persistent desk rather than more conversation threads.
Thrymvault is described as a workspace for pages, databases and boards, built and hosted on the publisher’s hardware. Agents can connect through an MCP server with 20 tools, a REST API or a command-line tool. Meyer says the agent guide is generated from the same tool catalogue as the server, so its instructions match the available tools.
In the proposed workflow, a topic card moves from the editorial board into a source-kit stage, then to drafting and publication. The editor can leave a comment mentioning an agent application to request a change; the agent is expected to respond in the same thread. Reports sit beside the relevant card, and a published story’s card ends with its live link. Moving a card is an instruction agents are meant to follow, Meyer says, rather than an automatic software trigger.
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 proposed setup addresses a practical coordination problem: agents can work in parallel, while chat messages and task status may be scattered across conversations. A board that keeps assignment state, editor feedback and draft links together could make it easier for an editor to see what is underway and intervene without reconstructing a story’s history from chat.
Attribution is also part of the design. Meyer says agent changes carry the name of the approved application they came through, displayed as “via” that application. He says agents enter only through applications approved by the workspace owner and cannot delete content. Those controls could help a newsroom review who or what made a change, though the source describes the intended system and does not provide an independent security assessment.
For readers, a record of who drafted a story, who requested a revision and which rule applied could make corrections easier to trace. That benefit depends on the workflow being followed and its records being accurate; the existence of an audit trail does not by itself verify a story’s sources or factual claims.
Products Worth Considering
LG UltraGear evo 52-Inch 5K2K 240Hz Curved Gaming Monitor
LG UltraGear 34G600A-B 34-Inch Curved WQHD Gaming Monitor, 160Hz
LG UltraGear 24-Inch Full HD IPS Gaming Monitor, 120Hz (144Hz Overclocked), 1ms MBR, G-SYNC Compatible, FreeSync, HDR10
As an affiliate, we earn on qualifying purchases.
From Chat Threads to Story Cards
Meyer’s account follows a specific newsroom use case, not a general release announcement for a finished product. He says his network, StrongMocha News Group, runs with AI agents and that Thrymvault is the workspace he is giving them. The article distinguishes between the design and what is already operating, but the supplied material does not give a full deployment inventory or independent product review.
The newsroom workflow assigns the editor responsibility for deciding which topics proceed. Agents can propose topics and perform tasks, while the editor steers work through card moves and comments. Rules that affect agent behavior, such as a headline policy, are recorded in a separate Decisions database, with a date, scope and link to the rule it changes. The rule itself remains where agents are instructed to read it.
The access model described by Meyer uses keys associated with a person and limited to one workspace, with permissions constrained by the application. The article says agents cannot delete anything. These details are attributed to the publisher’s guide and development log as of November 2026, rather than to an external audit.
“When I wanted to know where a story stood, I had to ask.”
— Thorsten Meyer, describing the newsroom’s workflow
Deployment and Safeguards Still Developing
The article does not establish how many agents or articles currently use Thrymvault, which parts of the workflow are already in routine production, or whether all the described controls have been tested under real-world conditions. Meyer says a second machine on standby and hourly backups are planned next steps, not current capabilities.
It is also unclear from the source how the system handles outages, compromised credentials, conflicts between simultaneous edits, or mistakes made by an approved agent application. The claim that each change is attributable to an application describes the design; it is not independent evidence that every action can always be identified or that an agent’s work is correct.
Standby Hardware and Backup Plans
Meyer identifies a standby machine and hourly backups as the next infrastructure steps. The source gives no launch date for those changes. He also says the article separates what is running today from features still being rolled out, but the supplied material does not provide a complete, dated checklist of deployment milestones.
For now, the described next stage is continued use and rollout of the shared desk: agents update story cards and reports, editors steer work through comments and decisions, and published cards retain links to the finished articles. Whether that process becomes the newsroom’s standard workflow, and how its reliability and access controls perform over time, remains to be reported.
Key Questions
What is Thrymvault?
Thrymvault is a workspace for pages, databases and boards that Thorsten Meyer says his group built and hosts on its own machines. It is intended to give editors and AI agents a shared place to coordinate newsroom work.
How do agents connect to the workspace?
Meyer says agents can use an MCP server with 20 tools, a REST API or a command-line tool. Access is through applications approved by the workspace owner.
Can an AI agent publish or delete a story?
The described workflow has agents prepare work and editors decide what proceeds. Meyer says agents cannot delete content; the source does not fully specify the publishing permissions for every application.
Are backups and a standby machine already operating?
No. Meyer identifies a second machine on standby and hourly backups as planned steps, not capabilities that are live at the time of his account.
Does an audit trail guarantee that an AI-written story is accurate?
No. A record of which application made a change can support review, but it does not independently verify sources, figures or claims in an article.
Source: ThorstenMeyerAI.com
