Neon now has realtime - Neon
Neon has launched Realtime in public beta, letting any SQL query stay live as the underlying data changes, with the sync running inside Postgres rather than in a separate service. The pitch is aimed at developers and coding agents who want reactive apps without polling or refetch logic.
- Any read-only, deterministic SQL query can be subscribed to, including joins, aggregates, and window functions, and the result updates as the data changes.
- Realtime runs inside Postgres, so there's no separate sync system to mirror or keep consistent with the database, and it's enabled per branch.
- It's free during the beta with no waitlist, and it's designed so it doesn't add load to the primary or keep a scale-to-zero database awake.
Neon Realtime is now in public beta. The core idea is that any SQL query you write can become a live query: subscribe to it, and the result updates as the underlying data changes, with no polling or refetch code. The pitch rests on a specific claim. Realtime has usually lived in a separate service that mirrors the database, and that mirror is where things go stale or break. Neon's version runs inside Lakebase Postgres itself, so there's no second system to keep in sync, and it supports scale to zero and branching. It's free during the beta with no waitlist. The team says it's built on what they learned building the Electric sync engine, but rebuilt for Neon's storage architecture, which they argue makes fully expressive queries (joins, aggregates, window functions, subqueries) feasible at enterprise scale.
The practical surface is a CLI command (`neon realtime subscribe`), a TypeScript SDK, and a React hook called `useLiveQuery` that re-renders a component when its query result changes. You define queries on the backend using raw SQL, Drizzle, or Kysely, then "seal" them so they can't be tampered with and can safely be handed to a browser or agent. Writes can be applied optimistically on screen and confirmed when Postgres commits the transaction, and TanStack DB and SSR are supported. The example in the post is a channel list with unread counts computed in a single SQL query, which is the kind of thing that normally takes a fair amount of glue code.
The article is also explicit about its audience: coding agents. The argument is that agents usually get CRUD right but botch realtime, largely because the sync layer sits outside the backend with its own permissions and no awareness of which branch it's pointed at. By keeping realtime inside Postgres and enabling it per branch, Neon wants agents to reach for it by default, with one command to turn it on. The post is promotional, and claims like "works with scale to zero" and "scales to the largest enterprise workload" come without benchmarks or customer numbers in this excerpt, so those are worth checking before relying on them. The team is asking people to try to break it and report findings on Discord ahead of general availability.