PowerSync

@arewa_coder @EOEboh 👋

@EOEboh You'll most probably need a Local database like sqlite and a sync engine. 🤷‍♂️

You're building an offline-first note-taking app Users need to store thousands of notes, search through them, and sync when back online. What's the actual tool for this job?

A few months ago I tested LLMs as a prototyping accelerator... and ended up going down the rabbit hole of adding support for Convex to PowerSync... and ended up actually shipping it 🙈 If your users are asking for your Convex app to work offline, I'd love to hear your thoughts on this. Notes 👇

We shipped initial support for Convex as a source database. "Adding PowerSync to a Convex app gives you offline-first behavior, optimistic updates, and partial sync."

We got our @powersync_ dashboard rewrite blogpost out. It took a while but at least we got it out before Bun could get their rust port post out 😅. Good read by @Manrich121. powersync.com/blog/rebuildin…

We did the same thing recently, migrated the powersync[dot]com website to Tanstack Smooth sailing

Kobie walks us through the new search functionality in the dashboard. Which features are we still missing?

Full thread: reddit.com/r/FlutterDev/c…

Evidence of how Sync Streams solve partial sync at scale - in this case a user is syncing 5B rows in their backend database to millions of rows in SQLite

Self-managed Private Endpoints are now available: - Connect your source database over private AWS networks - Zero public internet exposure - Fully self service - no need to contact support releases.powersync.com/announcements/…

More detail and link to the PoC: github.com/powersync-ja/p…

In this fun video, Steven shows us how he created an experimental @MikroORM integration. For existing MikroORM users, PowerSync provides partial sync at scale: keep on-device SQLite synced with your backend database. PowerSync is designed for offline, real-time, and reactive use cases. For existing PowerSync users, you might want to check out MikroORM as another great option. From what we can tell, the latest version probably works on React Native and Web too.

Repo: github.com/powersync-comm… The specific example is for React Native, but the approach generalizes to any of our supported SDKs

We often speak to customers that are syncing massive (1GB+) volumes into SQLite This is something we'll have better support for over time, but a good workaround for now is to use the PowerSync SQLite Seeder community project

Partial, consistent sync at scale is one of the hardest problems solved by PowerSync. We first launched with a declarative system called Sync Rules, which synced a user’s entire dataset to their device on first connect. This fell short in use cases where you didn't want to or need to sync everything up-front. So we built Sync Streams: a system where clients can still sync everything up front, or subscribe lazily to exactly what they need, when they need it, using parameters. Sync Streams has been soaking in the wild for the past few months and is now GA / Stable. Thanks to everyone who helped us put it through its paces.

Using our reactive watch queries, you don't need to build UI invalidation by hand. Available on all our SDKs. Great article by @InfoWorld promoting sync as a paradigm

"Local-first is not offline-first." This is a clear take on why the distinction matters, and includes a kind nod 🥰

@rodydavis 👍 What's your sentiment on the overall stability of OPFS in Chrome?

@powersync_ I do file blobs with OPFS and a POSIX metadata layer with indexedDB. That way I get super fast file/folder manipulation and can support symlimks while heavy ready/write blob streaming is done via OPFS and can even deduplicate.

Persistence on the web in prod?

Full discussion - github.com/rhashimoto/wa-…

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

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