Over the second half of last year, we focused intensely on the technical foundations of our work. That effort led to a set of infrastructure breakthroughs that reduced our server costs by more than an order of magnitude.

As a result, funds we had already set aside to cover hosting through 2026 will now cover our PubPub server costs for several additional years.

This means two immediate things:

1. Groups currently using PubPub can continue doing so without interruption beyond 2026.
2. We have a financially feasible path to supporting the server costs of PubPub indefinitely.

---

Sustaining This Work

While the funds we previously set-aside will sustain server costs for the next few years, they are not a permanent solution. To address this, we’ve established the PubPub Sustainability Fund.

This Fund reflects a new reality for PubPub: server costs are now sufficiently low, and ongoing maintenance sufficiently reduced, that a relatively modest pool of funding can keep PubPub online indefinitely. As long as the servers are funded, we are committed to keeping PubPub available.

Our goal is $300,000, which would cover core server and technical costs for the foreseeable future. Every donation, at any amount, strengthens PubPub’s ability to remain a shared resource for the public knowledge community.

---

PubPub and PubPlatform: Two Independent Paths

With a sustainable path now in view for PubPub, the previously planned migration to PubPlatform is no longer required; PubPub and PubPlatform will continue independently.

• PubPub: a full-stack, turn-key academic publication site builder. (The tool you’ve known and loved at www.pubpub.org.)
• PubPlatform: highly customizable, low-level infrastructure for building bespoke publication toolchains at a fraction of the cost of developing a brand-new application from scratch.

This separation reflects different use cases and different sustainability models, and it allows each project to evolve on its own terms.

Where the Cost Reductions Came From

The reductions we achieved came from three primary areas. We are happy to go into technical detail with anyone who is interested.

1. Maturation of the Ecosystem

Many features that were cutting-edge when PubPub was first developed — such as our multi-tenant routing, custom domain management, multi-tenant SSL certificate handling, CI/CD pipelines, and full-text search — now have mature, open-source alternatives or are available as built-in, no-cost services from major providers.

A concrete example is the widespread adoption of Let’s Encrypt. Native certificate automation allowed us to refactor certain pieces of our infrastructure stack such that we can move away from infrastructure providers that were significantly more expensive.

2. Better Database Tooling and Prior Investments

The tooling available for inspecting and diagnosing expensive PostgreSQL queries has advanced dramatically since the mid-2010s. Combined with our earlier investment in converting the codebase from JavaScript to TypeScript, this allowed us to identify costly queries and refactor them into functionally identical — but far less expensive — forms. This means we can run fewer servers and cheaper servers, while still providing equal or better site performance.

3. Deprecating Rarely Used Expensive Features

We also analyzed features that were expensive to maintain but rarely, or never, used.

As part of this refactor, one feature will be deprecated: step-by-step, keystroke-level edit history of documents for all pubs, indefinitely. While novel, this feature carried enormous storage and processing costs, and our data shows it was used by fewer than 0.1% of users. Versions, release history, and collaborative editing are what most communities rely on in practice.

For groups that wish to preserve this functionality, we can support it on a per-community basis and pass through the associated costs.

Improved Self-Hosting and Long-Term Resilience

This infrastructure refactor has also made self-hosting PubPub Legacy significantly simpler. While it is not yet fully independent of external services (such as AWS or Firebase), standing up local or managed instances is now feasible using standard Docker workflows. We are happy to support organizations interested in doing so and welcome community contributions that further improve the self-hosting experience.

This capability is also part of our long-term resilience plan. Should PubPub face financial pressure again years from now, handing off hosting to another organization would be far less expensive and far less technically complex than it once was. We intend to build a consortium of technically capable organizations to support this work — and potentially host mirrors — to ensure long-term access to PubPub content.

Acknowledging the Past Year

We know that many communities went through an uncomfortable process trying to decide if, when, and where to migrate following our June 2025 announcement. At that time, we could not guarantee this technical progress, and we did not want to offer assurances we could not fully stand behind. In reality, we moved as quickly as we responsibly could, but still our work required sustained effort over several months to ensure these improvements are real and durable.

Over the past 6 months we heard the same message repeatedly: there is nowhere else to go. The absence of clear alternatives for publishing diverse, non-traditional knowledge artifacts underscores the importance of ensuring PubPub remains available for current and future communities.

We are deeply grateful and extend a heartfelt thanks to everyone who helped us navigate the turbulence of 2025 and supported us as we found a new way forward.