🔍 Read the full analysis: Building A Shared Workspace For AI Agents With Thrymvault 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 StrongMocha News Group is building Thrymvault as a shared workspace where editors and AI agents can track newsroom work. The system is designed around approved applications, attributable changes and limits on agent permissions; some planned infrastructure, including a standby machine, is not live yet.
Thorsten Meyer says his StrongMocha News Group is building Thrymvault, a self-hosted workspace intended to let editors and AI agents manage newsroom work on shared boards and pages. The development addresses a problem Meyer encountered while coordinating nine articles through chat and a code-oriented task tracker: story status, editorial instructions and agent changes were difficult to see and track in one place.
In an article published November 13 on ThorstenMeyerAI.com, Meyer describes Thrymvault as a workspace for pages, databases and boards that his group operates on its own hardware. He says the interface is designed for agents as well as people, with access through an MCP server, a REST API or a command-line tool. The MCP server currently has 20 tools, according to the article.
The proposed newsroom workflow puts each story on an editorial board. An editor can move a card from “Topic” to “Kit” to signal that an agent should gather sources and prepare a source kit. An agent then drafts the article and updates its board entry with a status, draft link and open questions. Meyer says requests can be added as comments that mention the relevant agent application, keeping the instruction and response in the story’s thread.
Meyer says agents can access the workspace only through applications approved by a workspace owner, and that each change identifies the application used. The article also says agents cannot delete content. A key is limited to one workspace and has no more rights than its application allows, according to Meyer’s description. He distinguishes the current system from future plans: a second machine on standby and hourly backups are not live yet.
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 project is aimed at a practical coordination problem for teams using multiple AI agents: a chat conversation can contain instructions, but it may not provide a shared, durable view of a story’s status, source material and revisions. A board that keeps those items together could make editorial work easier to monitor and hand off, particularly when agents operate in parallel.
For newsroom accountability, Meyer’s design also makes the route of a change visible. A record that identifies which approved application made an edit, alongside comments and editorial decisions, could help an editor investigate or correct a mistake. That is a design goal described by Meyer, not evidence that the system has been independently tested or that it will prevent errors.
The approach reflects a wider operational choice: instead of relying only on a chat service or a third-party workspace, Meyer says the group runs Thrymvault on its own machines. That gives the team direct control of its hosting environment, while also making the team responsible for keeping the system available and protected. The article does not provide performance, security-audit or cost results.
Popular right now
TCL 65-Inch 4K UHD Smart LED TV (65S425)
Dyson V9 Motorbar Cordless Vacuum
TCL 55-Inch T7 Series 4K QLED Smart Google TV (55T7)
As an affiliate, we earn on qualifying purchases.
From Chat Threads to Story Cards
Meyer says the immediate impetus was a day in which his team steered nine articles using a chat window and a task tracker built for code work. He describes having to ask agents for updates, repeat headline instructions in chat and find that a completed explainer could not be shared as a link. Those are his account of the workflow problems that prompted the project.
Thrymvault is presented as a desk for that work rather than as a replacement for the publishing system. In Meyer’s described process, the board tracks the story from topic selection through research, drafting and scheduling, with the published link added at the end. Editorial choices that affect future work—such as a new headline rule—are recorded in a Decisions database, while the rule itself remains where agents access it.
The agent guide is published as both a plain-text page and a machine-readable manifest, Meyer says. Both are generated from the same tool catalogue as the MCP server so the written instructions match the available tools. The article identifies this as part of the system’s design and cites Thrymvault’s guide, integration documentation and development log as sources for its descriptions.
“That is the point where an AI-run newsroom outgrows chat.”
— Thorsten Meyer, describing the newsroom’s workflow
Rollout and Safeguards Still Developing
Meyer’s article describes a mix of existing capabilities and a workflow that is still being rolled out. It does not specify which parts of the newsroom process are already in routine use, how many agents or users currently rely on the workspace, or when the remaining features will be complete. The planned standby machine and hourly backups are explicitly not live.
The account is from the project’s creator and does not include independent testing or an external security review. Details such as the full permissions model, recovery process, uptime, data-retention practices and results from using the system at scale are not provided. The article also does not say whether Thrymvault is available to other organizations, or give pricing or a public product-release schedule.
Although Meyer says changes are attributable to approved applications and agents cannot delete content, the material does not explain how the system handles compromised credentials, mistaken edits or disputes over an agent’s actions. Those questions remain open; the described controls should not be taken as proof that errors or unauthorized access are impossible.
Backups and Wider Rollout Ahead
Meyer says the next infrastructure step is to add a second machine on standby and hourly backups. The source does not give a delivery date. He also describes features as being rolled out, but does not publish a timeline for completing the newsroom workflow or making Thrymvault available beyond his organization.
For now, the next meaningful indicators will be whether the planned backup setup goes live and whether the team can show how the workspace operates in regular editorial production. Further reporting would be needed to establish availability, reliability, security-review status and whether other newsrooms can use the system.
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 hardware. It is designed to let people and AI agents coordinate work in a shared environment.
How can AI agents use the workspace?
Meyer says agents can connect through an MCP server with 20 tools, a REST API or a command-line tool. Access is limited to applications approved by the workspace owner, according to his article.
What controls does Meyer say are in place?
The article says changes identify the application that made them, agents cannot delete content, and an access key is limited to one workspace and to the rights permitted for its application. These are claims in the project creator’s description; the article does not cite an independent security review.
Is Thrymvault’s backup system live?
No. Meyer says a second machine on standby and hourly backups are planned as the next step, but are not live yet. He does not provide a date for their rollout.
Can other organizations use Thrymvault?
The source describes how Meyer’s StrongMocha News Group plans to use the workspace but does not say whether it is available to other organizations, or provide a release schedule or pricing.
Source: ThorstenMeyerAI.com
