TL;DR
Threlmark has described a local-first architecture for its project management tool that uses plain JSON files on disk instead of a database, cloud service or user accounts. The company says the file layout itself acts as the API, allowing the UI, external tools and AI coding agents to read and write the same project records.
Threlmark has detailed a local-first architecture for its project management tool, saying the product runs as a Next.js app over plain JSON files on a user’s disk rather than a database, cloud backend or account system, a design choice meant to make project data inspectable, portable and usable by outside tools and AI coding agents.
The report from Thorsten Meyer AI says Threlmark stores its data by default under ~/.threlmark, with a manifest, dependency graph, project folders, per-project board files, one JSON file per card, handoff records, report drop zones and a human-readable roadmap file. The source material says the central design rule is that the disk layout is the interface shared by Threlmark’s UI and other tools.
The source describes two safeguards for file-based state. First, writes are atomic: the app writes to a temporary file in the same directory, syncs it and then renames it over the target file. Second, each work item is stored in its own file, while lane ordering is kept separately in board.json and reconciled when the board is read.
Threlmark also says many displayed metrics are derived from the item files rather than stored separately. Priority, stale age, cycle time, throughput and WIP counts are described as computed values, which the report says reduces the chance that the UI shows numbers that no longer match the underlying JSON records.
Why It Matters
The design matters because it treats project data as local infrastructure rather than as records held inside a remote service. If the architecture works as described, users could inspect cards with standard tools, back up the folder, version it with Git, sync it with third-party storage or let another program participate by writing files in the expected format.
The approach also targets a newer workflow in which AI coding agents receive tasks and report completion. According to the source material, Threlmark supports a closed loop in which a card can be handed off, an agent can return a report through a REST endpoint or by dropping a JSON report file, and a completed report can move the card to Done.
JSON file project management tool
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background
Threlmark is being positioned as a project tool for deciding what should be built next across multiple projects, not only for tracking which lane a card occupies. The source contrasts that with common board tools that focus on card status.
The architecture described in the report is part of a series on Threlmark. The article says the app uses Next.js and TypeScript, stores records as JSON on disk and is released under the MIT license. The report also mentions a static read-only demo path using seeded data and localStorage, as well as a personal Node deployment path, but it does not provide full deployment details in the supplied material.
“The on-disk layout is the API.”
— Thorsten Meyer AI report
“There is no server-of-record — the files are the record.”
— Thorsten Meyer AI report
“Every artifact is a file you can cat, diff, grep, commit.”
— Thorsten Meyer AI report
“A handoff is a first-class flow event.”
— Thorsten Meyer AI report
local-first project management software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What Remains Unclear
Several details remain unclear from the supplied source material. The report does not state how many users are actively using Threlmark, whether the architecture has been tested under heavy concurrent use, what security controls apply to local report ingestion, or how conflicts are handled when file sync tools change the same records across machines.
The source also describes the architecture and intended behavior, but it does not provide independent benchmarks, third-party audits or production incident data. Claims about portability, interoperability and restartability are attributed to the report.
disk-based project tracking app
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What’s Next
The next question is whether Threlmark’s file-contract model can hold up as users connect more external tools and AI agents to the same project folders. Readers should watch for public documentation, repository activity, installation guidance, examples of report ingestion and evidence from real multi-tool workflows.
file-based task management system
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What did Threlmark announce or describe?
Threlmark described a local-first architecture in which a Next.js app reads and writes plain JSON files on disk, with the folder structure serving as the shared interface for the app, outside tools and AI agents.
Does Threlmark use a database or cloud account system?
According to the source material, no. The report says Threlmark uses local JSON files rather than a database, cloud backend or user accounts.
How does the system try to avoid file corruption?
The report says Threlmark uses atomic writes: it writes to a temporary file in the same directory, syncs it and renames it over the target file. It also stores one card per file to reduce write collisions.
How can AI agents interact with Threlmark?
The source says an agent can receive a handoff and then report back through a REST endpoint or by dropping a JSON report file. A completed report can move the related card to Done.
What is still not confirmed?
The supplied material does not confirm adoption numbers, production reliability, full security behavior, sync conflict handling or independent validation of the architecture’s claims.
Source: Thorsten Meyer AI
