What Rocicorp Zero 1.0 Is and Why the Stable Label Matters
Rocicorp Zero 1.0 is a general-purpose web sync engine that pairs a client library with a server cache to keep browser state aligned with a primary Postgres database, giving web applications fast local reads while syncing authoritative data in the background for collaborative use cases. Zero 1.0 is the first stable release after nearly two years of development, more than 50 releases, thousands of commits, and hundreds of bug fixes, formalizing an API that Rocicorp now commits to support. Functionally, the update is minor compared with version 0.26.2, but the semantic version bump is important for teams that were waiting on a stable surface before building large-scale collaborative applications. A stable API reduces the risk of future breaking changes and lowers the maintenance overhead for long-lived projects, which is often a deciding factor for enterprise adoption.
Inside the Web Sync Engine: Caching, ZQL, and the Bring-Your-Own-Database Model
Zero’s web sync engine architecture centers on a zero-client library in the browser and a zero-cache service that keeps a read-only replica of a Postgres database. Queries are written in ZQL, Zero’s streaming query engine, and run against the local cache first so users see matching results in the next frame, while the engine synchronizes authoritative data from the server in the background. According to InfoQ, project founder Aaron Boodman describes Zero as “just a fancy cache” that cannot corrupt data because it sits in front of, rather than replaces, the source database. This bring-your-own-database model lets teams keep their existing Postgres setup, integrate Zero incrementally, and back it out if needed. For developers building collaborative applications, this design offers responsive UI behavior with fewer custom sync mechanisms, while still keeping business logic and data ownership in Postgres.
Supabase Integration: Schema Change Hooks and Safer Sync Workflows
Zero 1.0’s most visible functional upgrade is a schema change hook aimed at improving Supabase integration. Supabase does not support event triggers for publications, which makes it harder to detect when replication configuration changes. To work around this, Zero now watches for COMMENT ON PUBLICATION statements and treats them as a signal that a publication’s schema has changed. This hook helps keep the web sync engine aligned with the upstream Postgres replication setup that Supabase manages, reducing the risk of silent drift between schemas. For teams building collaborative applications on Supabase, the result is a clearer schema change workflow: update publications, let Zero detect the change, and maintain confidence that the cached replica and client queries remain in sync without custom glue code or manual reconfiguration steps.
Bug Fixes, Limitations, and Production Readiness for Collaborative Applications
Beyond the schema hook, Zero 1.0 includes fixes for corrupted litestream restores, improved error handling when change data capture is unhealthy, an IPC_CHANNEL_CLOSED race during shutdown, orphaned subscribers, and time and timetz SQL generation. These fixes matter for production deployments that depend on long-running sync processes and reliable recovery. At the same time, early reviewers highlight limits: at present Zero supports only Postgres, does not sync views, and lacks support for some column types such as array. Critics also point to a 718 KB uncompressed (232 KB gzipped) client bundle and the absence of server-side rendering, which remains on the roadmap. Even with these constraints, the stable release signals that Zero is ready for more serious collaborative applications where latency, offline-friendly behavior, and predictable API evolution are more important than broad database coverage.
What Zero 1.0 Means for Web Developers and Enterprises
For web developers, the jump to Zero 1.0 means new projects can be planned around a stable web sync engine API designed for collaborative applications, rather than an evolving alpha surface. The upgrade path is simple, with no breaking changes from 0.26.2, which eases adoption in existing codebases. Teams coming from older versions still need to review historical release notes, including changes such as custom mutator types in 0.23, and follow the project’s schema deployment guidance to avoid downtime during migrations. In the wider sync ecosystem, Zero now stands alongside tools like ElectricSQL, InstantDB, PowerSync, Convex, and Jazz, but differentiates by focusing on Postgres and a bring-your-own-database cache model. For enterprises that want responsive collaborative experiences without abandoning their current databases, Zero 1.0 marks a credible, stable option worth evaluating.
