The Cloud Promise and Its Limits
Cloud-native applications offered a compelling trade: give your data to our servers, and we will make it available everywhere, back it up automatically, and let you collaborate with others in real time. For many use cases, this trade made obvious sense. Email, shared documents, project management tools — these benefit enormously from centralised infrastructure. The model spread rapidly because it worked well for the problems it was initially designed to solve.
The problems emerged more slowly. Cloud applications require a working internet connection. They require the continued existence and interest of the companies hosting them. They require trust that those companies will not change their pricing, terms, or privacy practices in ways that harm users. And they create a relationship in which the user is dependent on infrastructure they do not own or control.
These dependencies were always present, but they have become more visible as cloud-hosted services have shut down, changed business models, or simply degraded over time. A generation of users now has the experience of losing access to years of notes, photos, or documents when a service changes direction.
What Local-First Means
Local-first software prioritises the user's own device as the primary location for their data. The application works fully offline. If synchronisation exists, it is a feature layered on top of local storage rather than a requirement for basic functionality. The user's data is in a format they can access, back up, and potentially move to another application.
This is not simply a return to the desktop software of the 1990s. Modern local-first applications use conflict-free replicated data types and other techniques to enable real-time collaboration between multiple devices without requiring a central server to arbitrate conflicts. The technical challenges that made collaboration hard without a central server have largely been solved — the solutions have just not yet reached mainstream adoption.
Applications like Obsidian and Logseq demonstrate that sophisticated software can be built on local-first principles. They store data in open formats — often plain text or simple structured files — that the user can read, edit, and migrate independently of the application itself.
The Longevity Argument
The most compelling case for local-first software is durability. A document stored as a plain text file on your device will be readable in thirty years. A document stored in a cloud application's proprietary format is readable only as long as the application continues to exist and remains accessible. The history of software is littered with formats and services that seemed permanent and are now inaccessible.
This matters more for some types of data than others. Notes, journals, research, creative work — these are things people want to keep for years or decades. The cloud is well suited to things that need to be shared right now; it is poorly suited to things that need to be kept reliably across long time horizons.
The Tradeoffs
Local-first software is not strictly superior. Real-time collaboration between many users remains technically harder without central infrastructure. Automatic backup requires deliberate setup rather than being a default feature. The argument for local-first is not that it is always better, but that it should be more widely available as an option — particularly for personal data that users want to own, control, and keep.
The developers building in this space argue that the cloud-native paradigm has crowded out local-first alternatives not because local-first is technically inferior, but because the business model of cloud software — subscriptions, data leverage, lock-in — creates financial incentives that local-first software does not. As users become more aware of what cloud dependency costs them, the demand for alternatives is growing.