🔍 Read the full analysis: Why I Gave My AI Agents A Shared Desk 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 StrongMocha News Group is building Thrymvault as a shared workspace for AI agents and editors, replacing chat-and-task-tracker coordination with story cards, reports and traceable changes. The system uses owner-approved applications and limits agent access, but some planned infrastructure, including standby hardware and hourly backups, is not live yet.
Thorsten Meyer says his StrongMocha News Group is developing Thrymvault, a self-hosted workspace where editors and AI agents can track newsroom assignments, drafts and decisions in one place. The project follows a day when the group steered nine articles through chat and a code-focused task tracker, leaving Meyer to ask agents for updates and making it difficult to share completed work as a link.
Meyer describes Thrymvault as a workspace for pages, databases and boards, built and hosted on the group’s own machines. Agents can access it through an MCP server with 20 tools, a REST API or a command-line tool. The agent guide is generated from the same tool catalogue as the server, Meyer says, so the instructions agents receive match the available tools.
In the proposed newsroom workflow, an editor or agent places a topic on a board. The editor decides whether it advances; moving a card to a “Kit” stage signals the editorial agent to collect sources, check figures and prepare a source kit. During drafting, agents are expected to update their status and attach a report and draft link to the relevant card. Editors can request revisions in the card’s thread by mentioning the approved agent application.
Meyer says changes are attributed to the application used, rather than appearing anonymously, and agents cannot delete material. Each agent application requires workspace-owner approval. In the described setup, an access key has the rights of the person who created it, applies to one workspace and cannot grant more permissions than the application allows. The article distinguishes this design from what is already running: a standby machine and hourly backups are planned, not live.
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 Workflow for AI Drafts
A shared board could make parallel agent work easier for an editor to supervise than separate conversations. Story status, source reports, revision requests and publication links would sit alongside the assignment, giving people a place to see what is underway without asking agents for an update in chat. Meyer’s account describes an operational proposal, not independent evidence that the system has improved accuracy, speed or output quality.
The design also addresses accountability. When a change is linked to an approved application and editorial decisions are recorded, a newsroom has a clearer trail to consult if a draft or instruction needs correction. That does not establish that an article is accurate: editors still need to check sources, figures and final copy. The practical significance depends on whether the planned workflow is implemented consistently and whether its records remain useful as the volume of work grows.
Products Worth Considering
Logitech MX Keys S Combo Wireless Keyboard and Mouse, Graphite
Apple MacBook Pro 16.2-inch with M5 Pro, 18-core CPU, 20-core GPU, 48GB Memory, and 1TB SSD (Space Black)
Ledger Stax Crypto Wallet with Touchscreen
As an affiliate, we earn on qualifying purchases.
From Chat to Story Cards
Meyer’s account is grounded in the group’s own newsroom experience. On one day earlier in November, he says, the team steered nine articles using a chat window and a task tracker built for code work. He describes having to ask agents about progress, enter headline changes in chat and rely on the correct agent to notice them. A completed explainer page could not be shared as a link.
Thrymvault is presented as the group’s response. Meyer compares its general form to Notion, while emphasizing self-hosting and restricted agent access. The board-and-thread model is meant to keep instructions with each story rather than letting them scroll away in a conversation. Decisions that change how agents work, such as headline rules, are to be recorded in a separate database with a date, scope and link to the rule itself.
“When I wanted to know where a story stood, I had to ask.”
— Thorsten Meyer, describing the newsroom’s earlier workflow
What Is Running So Far
The account does not provide independent testing, usage figures or a detailed deployment schedule for Thrymvault. It describes the intended newsroom workflow and identifies some components, including the agent server and access rules, but does not specify how many agents or assignments are currently using the full system. Standby hardware and hourly backups are still planned, and the article does not say when they will be available.
It is also unclear how the workspace handles mistakes such as an agent acting on an outdated instruction, whether editors can restore or compare earlier versions, or how the group will measure whether the desk improves its reporting process. Meyer’s description should be read as an account of his own organization’s design and rollout, not as a verified assessment of newsroom performance.
The Next Infrastructure Steps
Meyer says the next infrastructure step is a second machine on standby, paired with hourly backups; both remain in development rather than live. The newsroom’s stated workflow would also carry assignments through source gathering, drafting, editorial revisions and scheduling, ending with a published link on the board.
The article gives no dates for those rollout milestones. Readers can look for updates on whether the backup setup becomes operational, how broadly agents use the board and whether the newsroom publishes evidence about the system’s reliability or effect on editorial work.
Key Questions
What is Thrymvault?
Thrymvault is a self-hosted workspace for pages, databases and boards that Meyer says his group is building to coordinate editors and AI agents.
How would agents use the shared desk?
Agents would use a board to update assignment status, attach draft links and reports, and respond to revision requests in a story’s thread. Access is through approved applications, using MCP, a REST API or a command-line tool.
Can AI agents delete newsroom material?
Meyer says agents cannot delete anything. He also says their applications require workspace-owner approval and that access is limited by the permissions attached to the key and application.
Is the backup system already in place?
No. Meyer says a second machine on standby and hourly backups are planned but not live. The article does not give a date for their deployment.
Does the workspace verify that articles are accurate?
The described workspace records assignments, reports and changes, but the source does not claim that it independently verifies articles. Editors still need to review sources, figures and drafts.
Source: ThorstenMeyerAI.com
