iiinova — where your digital self keeps evolving

Storage version 4 and S3 are rolling out automatically to Free plan instances. Other plans can opt in, with S3 enabled separately. The upgrade involves a one-time full client re-sync. Rollout details and setup instructions:

100

That reduces load on the storage database, improving initial sync speed and stability, especially with large datasets or many clients connecting at once. It also increases capacity for concurrent clients.

100

Bucket storage on S3 When enabled, larger batches of sync data move from the instance’s storage database to S3. The Service reads them from S3 and streams them to clients.

100

Available for MongoDB sources using Sync Streams. Support for Postgres and other source databases is planned.

100

With storage version 4, most Sync Streams deploys reprocess only the streams you’ve added or changed. Clients only re-download data for those affected streams. Some changes still require a full reprocess, but routine config updates become much lighter.

100

Incremental Reprocessing One of the biggest pain points with PowerSync was that even a small change to your sync config required re-replicating everything from scratch, and clients had to re-sync all their data.

100

PowerSync @powersync.com · 07/09/2026

PowerSync Service 1.26.0 is out, introducing storage version 4 with two big improvements: - Incremental reprocessing for MongoDB - S3 object storage for bulk sync data TL;DRs below 👇

100

Large attachment transfers just got easier in our JavaScript SDKs: you can now avoid buffering entire files in JS memory

100

PowerSync @powersync.com · 04/09/2026

Our Node.js SDK is now GA, after an extensive ~13 month soak :) V1.0.0 comes with more ergonomic APIs, simpler setup, and a bunch of new features. Release notes: releases.powersync.com/announcemen... Thanks to everyone who reported bugs and feedback during the beta!

releases.powersync.com

Node.js Client SDK v1.0: Now Stable

The PowerSync Node.js client SDK is now stable with the release of `@powersync/node@1.0.0`. Since the beta it moved to the v2.0 JavaScript API, added streaming attachments for large files, and reached feature parity with our other SDKs.

000

PowerSync @powersync.com · 03/09/2026

We’ve received several customer questions about bucket counting in Sync Streams. We’ve expanded our docs to explain how counts are calculated and added a troubleshooting guide for bucket count and parameter query result limits: docs.powersync.com/sync/stream... docs.powersync.com/sync/advanc...

docs.powersync.com

Reducing Bucket Count - PowerSync

Diagnose a high bucket count, reduce the number of buckets a user syncs, and raise the per-user limits when needed.

010

PowerSync @powersync.com · 01/09/2026

"We've been getting high PowerSync bills for the past few months", a customer told us. In response, our FDE team created a guide on how to reduce usage and lower your PowerSync Cloud bill: docs.powersync.com/resources/u...

docs.powersync.com

Reducing Usage - PowerSync

Practical ways to lower data synced and data hosted usage on PowerSync Cloud.

000

PowerSync @powersync.com · 31/08/2026

Custom roles are now self-serve in the PowerSync Dashboard on Team and Enterprise plans. Scope what an org member can view, deploy, and manage, including down to specific projects and instances. Docs: docs.powersync.com/tools/power...

000

PowerSync @powersync.com · 28/08/2026

PowerSync Cloud is now infrastructure-as-code. We shipped an official Terraform provider: declare projects, instances & sync config ➡️ terraform apply ➡️ done Docs: docs.powersync.com/tools/terra... AWS ECS example: github.com/powersync-c...

github.com

GitHub - powersync-community/terraform-self-host-aws: A reusable Terraform deployment of a self-hosted PowerSync stack on AWS

A reusable Terraform deployment of a self-hosted PowerSync stack on AWS - powersync-community/terraform-self-host-aws

100

In plain English: NORMAL/OFF = faster but may lose latest writes; FULL/EXTRA = safer but slower.

100

If the app needs stronger guarantees, it can turn durability back up with PRAGMA synchronous=FULL or EXTRA. Then the VFS flushes more aggressively so recent commits are much more likely to survive crashes, but write performance drops.

100

So the tradeoff is simple: faster writes now, weaker crash protection for newest data. This is often fine for local-first apps where recent changes can be retried, re-synced, or recreated from server state.

100

Why the speedup is big: forcing storage flush on every tiny write is expensive. That flush is often the slowest part of the commit. By skipping it on a hot path, the VFS removes a lot of overhead, which especially helps with workloads with many small transactions.

100

In default mode, commit means "this transaction is recorded in WAL and the database stays valid," not "the browser has definitely flushed this to storage." If the tab crashes or the device loses power, the app may lose the most recent writes, but the database should not come back corrupted.

100

Key idea: keep the database correct even if newest writes are not fully durable yet.

100

It does this by not insisting every commit hit the disk immediately. Instead of updating the main database file right away, it writes changes into side log files first. That is cheaper, so small write transactions finish much faster.

100

PowerSync @powersync.com · 01/05/2026

The new OPFSWriteAheadVFS for wa-sqlite not only brings concurrent reads to SQLite on the web, but also makes write transactions faster!

110

We're at @stirtrek! Come say hi, grab a free tshirt, enter a giveaway for a Meta Quest 3S, and chat about building fullstack AI applications

000

PowerSync @powersync.com · 30/04/2026

PowerSync is proud to be sponsoring this year’s @stirtrek! And our own @devagrawal09 will be speaking on building Fast, Realtime, and Offline-First Apps with Sync Engines x.com/stirtrek/st...

000

PowerSync @powersync.com · 29/04/2026

📝 supports write-ahead logging (WAL) ⚡ moves data faster than IndexedDB ✍️ commits writes faster than IndexedDB 📁 lets you copy, import, export, and rename DB files with OPFS? Say hello to OPFSWriteAheadVFS, a new virtual file system option for wa-sqlite!

100

PowerSync @powersync.com · 29/04/2026

How would you like a persistent browser database that: 💾 stores data durably in the browser 🪶 avoids Asyncify time and size overhead 🌐 skips COOP/COEP header requirements 🗂️ works across multiple tabs and workers 👀 supports simultaneous read transactions

100

PowerSync @powersync.com · 27/04/2026

If your AI app needs durable runs, resumable streams, background jobs, and shared history to be reliable, you're already partway to multiplayer. The next step is making that state collaborative so that your agents can align with not just a single user, but an entire team.

100

Result: browser SQLite that behaves much more like a normal SQLite WAL database under concurrent reads.

100

To keep multiple tabs and workers in sync, connections send lightweight messages over BroadcastChannel after commits and checkpoints. They also use navigator.locks to advertise which range of WAL history each connection still needs.

100

So the model is really "many readers, one writer" rather than "everyone waits for everyone" or “writers block readers”.

100

A writer uses separate a write lock, so it can append new changes to log without blocking readers that already have a read snapshot open. Readers do not block the writer, and the writer does not block readers.

100

That is the key reason simultaneous reads work: active readers are not forced onto moving targets every time a writer commits.

100

When a read transaction starts, the connection effectively "freezes" its view of the database. It will keep reading from the same snapshot even if another tab commits new writes in the meantime.

100

This gives each reader its own consistent snapshot without needing global shared memory.

100

Each tab or worker keeps its own local map of where the latest version of each page lives. When the read query runs, it first checks whether the page has a newer version in the write-ahead log; if not, it reads the page from the main DB file.

100