Today we launched the beta of Sync Streams, our new partial sync architecture that is the successor to legacy Sync Rules.
Sync Streams is a flexible format that lets you define exactly which data from your backend can sync to each client using simple SQL-like queries.
The PowerSync CLI just went open-source and received an upgrade!
- Support for self-hosted and Docker workflows
- Browser based login flow
- Validate Sync Stream definitions and migrate from Sync Rules
- Plugin system
- ...and much more!
PowerSync now has Agent Skills!
Claude Code can now build real-time and offline-first apps for you.
Give your favorite coding agents the ability to ship instant UX and reliable data sync without learning a new database language!
$ npx skills add powersync-ja/agent-skills
I wrote another blog post!
Apparently building apps that work offline sounds scary to people
Turns out it's not that scary, just a matter of making the right tradeoffs at the right time
Want to make your app work while offline?
Great news! All you need is your existing backend database and @powersync.com
with some @tanstack.com DB for additional seasoning!
๐๐ป๐ป๐ผ๐๐ป๐ฐ๐ถ๐ป๐ด ๐๐ต๐ฒ ๐ฃ๐ผ๐๐ฒ๐ฟ๐ฆ๐๐ป๐ฐ ๐๐ ๐๐ฎ๐ฐ๐ธ๐ฎ๐๐ต๐ผ๐ป ๐ค๐งโ๐ป
We are hosting a virtual hackathon where the challenge is to build innovative AI-powered software using PowerSync as a sync engine.
dm
the focus should still be on the framework user
we just need to recognize that a lot of the framework users are gonna be ai agents
new saas provider - powersync can be self hosted, but yes that is one additional thing to deploy/monitor. fully worth the benefits imo, as it also means you never have to worry about putting a redis cache in front of your database to speed up reads.
defining a schema - powersync makes this one really easy, it can autogenerate the schema for you by looking at the data coming from your backend db
new backend adapter - yes technically it's a "new thing", but in practice you basically move the mutation and auth logic scattered throughout the codebase into a single backend endpoint that handles syncing, so in the process of adding a backend adapter you also remove a lot of the existing logic
wrapping localStorage in a signal only gets you the ability to read data offline, nothing else
and even then if you want to guarantee consistent reads you almost have to rebuild an entire database on top of localstorage, at which point, just use sqlite or tanstack db
> no error handling
that sounds less like "optimistic UX" and more like "terrible UX"
i describe a few in this blog post. real time is only one of the benefits of sync engines. www.powersync.com/blog/unleash...
what's the difference between Optimistic UX and Optimistic UI for you?
I agree that adding so much infra complexity just for optimistic UI is not worth it. you can imagine a much simpler sync engine that doesn't require additional infra and provides you instant local reads and optimistic ui.
i know.
sync engines eliminate the need to manually manage the cache lifecycle. you can cache stuff without a sync engine but only if you own the invalidation logic.
so offline access then
still, I'd use a sync engine even if it did not come with realtime updates or offline availability, because the stuff I care about the most is the cache logic and optimistic updates
can't imagine what you're going through but know that you are one of the brightest members of our community and we got your back
don't rewrite all the reads then, migrate incrementally, piece by piece
you can do that when a sync engine integrates with your existing tools
would you say real time updates are the main benefit of a sync engine?
"collaboration" sounds the same as "realtime updates"
i have a few others that i described here www.powersync.com/blog/unleash...
interesting
i think of sync engines as removing a layer instead of adding
www.powersync.com/blog/unleash...
the argument against sync engines would be that the rearchitecture effort is not worth it, and the end result is not simpler or faster and does not meaningfully improve DX and UX
and if you're familiar with sync engines you know that isn't the case
can you elaborate more on "it is sending you the world"
so really it comes down to the payoff - whether moving to a sync engine from a custom caching/rpc setup brings long term benefits
i would argue that it does - the architecture you end up with is much simpler than either setups you described which should improve velocity in the long term
so you're saying that someone who build their custom bespoke sync engine will need to rearchitect things to switch to an off the shelf sync engine.
pretty sure this is true for library or framework worth adopting, nothing comes for free
The argument for the client-side machinery is that without it you would be building very similar but bespoke machinery for your application instead of putting engineering effort into the business and application logic
like cache/query invalidation or optimistic UI logic
are real time updates the primary benefit of a sync engine in your perspective?
do you have some examples of this? very curious as to what sort of rearchitecture needs to happen to incorporate a sync engine in your experience
so what makes you say that you would almost never use a sync engine?
clearly they provide a lot of ux and dx benefits
